酒店客房的超额预售

No abstract available

No abstract available

No abstract available

The application of the overbooking process for different segments of the hotel services market requires the study of factors and prerequisites for influencing demand. The main problem faced by hotels when using overbooking is that, despite a set of booking requests that exceeds the total number of rooms, there is a lack of room capacity in the intervals between bookings. Hotel owners and managers, including those with a limited capacity, are increasingly interested in revenue management methods, but not everyone has managed to implement a system of Revenue Management, which is an integral part of effective management. On the one hand, a revenuemanager can optimize profitability and profitability not only from room sales, but also from other hotel revenue centers; on the other hand, knowing (analyzing) the habits of regular guests, it is possible to increase their loyalty by offering a personalized approach to providing special conditions for hotel services. Determining the potential of bookings and their impact on the total revenue of a hotel is based on an arsenal of data on needs in the general economic aspect, which form the interests, expectations, and priorities of consumers in choosing a hotel service. From the perspective of market relations, it is the needs of consumers that are the basis for the formation of consumer demand for hotel services. In the practical context of distribution of hotel services, the main dispute in the formation of a room reservation guarantee system is reflected in the content of compensation processes for reasons of no-show or early departure. In accordance with this statement, the fundamental content, prerequisites and benefits of using overbooking for the guest are the flexible consumption, at the level of the hotel enterprise – it is to ensure the process of managing the occupancy as a set of separate areas of planning and maximizing revenues. However, it is worth noting that the increasing application of overbooking depends on individual decisions of the hotel company and requires differentiation of the hotel service, which, in particular, involves expanding the price plan, introducing new methods in the process of their promotion and transforming distribution. The problem is that the level of overbooking is determined primarily by the method of trial and error and can lead either to resale of rooms and forced relocation of guests or to an unreasonably large quantity of vacant rooms and a lack of revenue.

No abstract available

Abstract We develop a hotel revenue management optimization method in an environment where market segment prices are optimized via demand curves ahead of a planning horizon. This new method simultaneously optimizes overbooking levels and allocation (of capacity to market segments) levels, as opposed to the traditional sequential approach. We test our method against the reference in a simulation of a hotel reservation system that has all the functionality of a real-world revenue management system: the estimation of true demand from censored demand; different market segments with different demand patterns; price elasticities; varying propensities to stay certain lengths of time; short- and long-term forecasting with periodic reoptimization of all forecaster parameters; explicit optimization of market segment prices based on estimated demand curves; and optimization routines for overbooking and allocation. A walkthrough of this simulation was performed by the revenue management staff at a major hotel. This simulation has been scaled down to permit extensive experimentation. Our new method outperforms the reference method by an average of 20.2% with respect to nightly net revenue. The improvement is much larger in situations where demand is more saturated. Our new method takes less than two minutes of computing time from a cold start on a realistically sized problem, which is sufficiently fast for hotel managers who want the capability of rerunning the algorithm many times during the course of a day.

No abstract available

No abstract available

Booking cancellations have a significant impact in demand-management decisions in the hospitality industry. These cancellations often limit the production of accurate forecast which is a significant tool in terms of revenue management performance. In order to curb these booking cancellation problems, hotels tend to employ rigid cancellation policies and overbooking strategies which could be detrimental to revenue generation and reputation of the hotel. This study is aimed at creating a machine learning model that could, with very high accuracy and precision, predict hotel booking cancellations. The dataset used is based on the individual bookings drawn from a hotel reservation system from a resort hotel in Portugal. The model was built using random forest algorithm with the dataset being split into 80% for the training set and 20% for the test dataset. By addressing booking cancellation prediction as a classification problem in data science context, the findings showed that it is possible to build models for predicting booking cancellations with an accuracy result of over 88%. This allows hotel managers accurately predict net demand and build better forecasts, improve cancellation policies, define better overbooking strategies and hence come up with more realistic and profitable pricing and resources allocation strategies. Keyword: Machine Learning, Hotel Booking Cancellation, Random Forest Algorithm, Prediction. CISDI Journal Reference Format Afolorunso, A.A., Aimuel, B.E., Abiodun, A.O., Adesina, A.O. & Ajagbe, S.A. (2025): Development of a Machine Learning–Enabled Decision Support Framework for Hotel Booking Cancellation Prediction. Computing, Information Systems, Development Informatics and Allied Research Journal. Vol 16 No 3, Pp 23-36. Available online at www.isteams.net/cisdijournal. dx.doi.org/10.22624/AIMS/CISDI/V16N3P2

Hotel booking cancellations pose significant challenges to the hospitality industry, affecting revenue management, demand forecasting, and operational efficiency. This study explores the application of machine learning techniques to predict hotel booking cancellations, leveraging structured data derived from hotel management systems. Various classification algorithms, including Random Forest, XGBoost, and LightGBM were evaluated to identify the most effective predictive model. The findings reveal that XGBoost model outperforms other models, achieving F2-score of 0.7897. Key influencing factors include deposit type, total number of special requests, and marketing segment. The results underscore the potential of predictive modeling in optimizing hotel revenue strategies by enabling proactive measures such as dynamic pricing, targeted customer engagement, and improved overbooking policies. This study contributes to the ongoing advancements in data-driven decision-making within the hospitality industry, offering insights into how machine learning can mitigate financial risks associated with booking cancellations.

The increasing uncertainty in travel has resulted in elevated cancelation and no-show rates across many aspects of travel, elevating the importance of overbooking practices. Overbooking helps address travel uncertainty by accepting reservations beyond available rooms but may result in walk or re-accommodation costs if all (most) of these reservations materialize. Walk costs are not homogeneous across all customer types, with costs potentially different for loyal (branded, direct) versus non/less-loyal (third-party-intermediated) guests. We formulate an optimal overbooking model with class-dependent walk-out costs for a hotel with two classes of reservations—loyal members with higher walk-out costs, and nonmembers with lower walk-out costs, but with each class at the same room rate. We embed a dynamic walk-out model, one where guests may be proactively walked, that is, walked while rooms still available, into an overbooking model. The joint model determines optimal walk-out decisions to minimize expected walk-out costs while also determining optimal overbooking levels. We investigate how class-dependent no-show rates and walk-out costs impact optimal walk-out decisions and optimal overbooking levels. We find that changes in the no-show rates for a customer class only impact the overbooking levels of the related class whereas changes in class-specific walk-out costs impact all customer class overbooking levels. We offer managerial insight into a proactive and strategic walk-out policy for the lodging industry, aiming to achieve optimal overbooking levels.

No abstract available

Booking cancellations in the hospitality sector not only result in revenue loss, but they also have an impact on billing and stock or resource allocations and budgeting. The hospitality industry’s demand management strategies are greatly influenced by booking cancellations. The ability to produce accurate projections, a crucial tool for revenue management success, is constrained by cancellations. According to 2016 data, the majority of cancellations are made through OTAs, with Booking.com taking the top spot at 57% and Expedia at 26%, compared to official hotel websites, which have an average cancellation rate of 14%. Hotels often employ strict cancellation rules and overbooking techniques to avoid the issues brought on by cancelled reservations, but these tactics can also negatively affect income and reputation. The outcomes enable hotel managers to precisely forecast net demand, create better projections, enhance cancellation procedures, specify better overbooking techniques, and employ more forceful pricing and inventory management tactics.

No abstract available

In the hospitality industry, demand forecast accuracy is highly impacted by booking cancellations, which makes demand-management decisions difficult and risky. In attempting to minimize losses, hotels tend to implement restrictive cancellation policies and employ overbooking tactics, which, in turn, reduce the number of bookings and reduce revenue. To tackle the uncertainty arising from booking cancellations, we combined the data from eight hotels’ property management systems with data from several sources (weather, holidays, events, social reputation, and online prices/inventory) and machine learning interpretable algorithms to develop booking cancellation prediction models for the hotels. In a real production environment, improvement of the forecast accuracy due to the use of these models could enable hoteliers to decrease the number of cancellations, thus, increasing confidence in demand-management decisions. Moreover, this work shows that improvement of the demand forecast would allow hoteliers to better understand their net demand, that is, current demand minus predicted cancellations. Simultaneously, by focusing not only on forecast accuracy but also on its explicability, this work illustrates one other advantage of the application of these types of techniques in forecasting: the interpretation of the predictions of the model. By exposing cancellation drivers, models help hoteliers to better understand booking cancellation patterns and enable the adjustment of a hotel’s cancellation policies and overbooking tactics according to the characteristics of its bookings.

Cancellations in bookings show a negative impact in the hospitality industry field while making management decisions. To prevent the negative impact of the cancellations, a lot of policies are implemented along with few overbooking techniques, this, in turn, can severely damage the income and reputation of that particular hotel. To prevent this situation, machine learning models have been developed. These models use previous data from the hotel and then it gets trained and predicts if the particular booking would get cancelled or not. Two hotels namely Resort hotel and City hotel have been considered, and then ML models are used to predict how particular actions taken by the hotel management shows impact on the hotel revenue and cancellations. This, in turn, makes management rethink about policies and their decisions. The Ml models will help management to predict the number of cancellations that may happen.

This paper presents an application of online booking data, comprised of big data crawled from a hotel booking website to hotel revenue management. It is important to build a quantitative revenue management method for online hotel booking systems incorporating overbooking strategies, because of increasing numbers of bookings through online booking websites and last-minute cancellations, which cause serious damage to hotel management. We construct a quantitative overbooking model for online booking systems combined with customers' choice behaviors estimated from the data. Firstly, we present the overbooking model for online booking systems. Secondly, we estimate the choice behaviors of the customers from the online booking data by a discrete choice model. Thirdly, combining the estimated discrete choice model with the theoretical overbooking model, we investigate the expected sales maximization problem where we numerically solve the optimal overbooking level and room charge. Finally, we provide numerical examples of the optimal overbooking strategies and room charges using online booking data of two major luxury hotels in Shinjuku ward, Tokyo. This method, which utilizes online booking data available by crawling from booking websites, helps hotels obtain an optimal room charge and overbooking level maximizing the expected sales.

No abstract available

No abstract available

No abstract available

Overselling is commonly adopted in the travel and hospitality sectors where a good or service is sold in excess of actual supply. We examine the impact of consumer foresight on efficient overselling when there are two dimensions of uncertainty, namely, early consumers are uncertain about their service valuations and the seller is uncertain about late demand arrival. We show that when consumers are naïve and have no foresight to anticipate future events, the seller resorts to the use of partial refunds and involuntary cancellation when the mandatory compensation for seller‐initiated cancellation is low, resulting in efficiency loss. When consumers are sophisticated and have perfect foresight on future events, efficiency is achieved when the seller sells the entire capacity in advance and relies solely on voluntary cancellation to re‐sell units when late demand warrants it. Refund complements overselling both by improving allocation efficiency in involuntary cancellation and by mitigating the cost of overselling when consumers have limited foresight. Unlike the social planner, the seller may suffer from consumer foresight. Our findings pinpoint the mandatory compensation in involuntary cancellation as a strategic tool for the social planner to tilt the seller's preference in the seller‐initiated cancellation policy to achieve efficient overselling.

No abstract available

No abstract available

No abstract available

There are a variety of revenue management systems that require making pricing or availability decisions for unique resources. For example, lodging marketplaces, boutique hotels, and bed-and-breakfasts offer unique rooms, apartments, or houses. Matching platforms for freelancers recommend differentiated workers with unique characteristics. When managing unique resources, one has to keep track of the availability of each resource at each time point in the future. Moreover, if the customers substitute between different resources, then the pricing and availability decisions for all resources become interdependent. Thus, it can be challenging to find good policies to make pricing or availability decisions. In “Revenue Management with Heterogeneous Resources: Unit Resource Capacities, Advance Bookings, and Itineraries over Time Intervals,” Rusmevichientong, Sumida, Topaloglu, and Bai consider revenue management problems when unique resources are requested for use over intervals of time under advance reservations. Using the interval structure of resource requests, they give policies with performance guarantees.

No abstract available

To maximize their revenues and protect their market share against traditional competitors (e.g. formal lodging businesses) and disruptive business models (e.g. Airbnb), the lodging industry increasingly relies on technology in various operations. However, the extent to which hotels adopt technology innovation in their revenue management (RM) operations, as well as the benefits of and barriers for digitalization, remains unclear. Moreover, the possible impacts of digital transformation on the future of revenue managers’ professions have been largely overlooked in previous studies. Drawing on qualitative data collected through 23 semistructured interviews with revenue managers in luxury and upscale hotels across Jordan, the findings suggest that RM is going through digital transformation with different levels of sophistication. While acknowledging the benefits of digital transformation in saving time, supporting the decision-making process, and yielding more revenues, the high cost of RM software emerges as a key barrier for digital transformation. The findings also reveal that the automation of various manual heuristics in RM is far from being possible, and therefore, digital transformation is unlikely to pose a threat to the future of the RM profession.

In the context of hotel revenue management, dynamic pricing plays a crucial role in maximizing revenue while maintaining a delicate balance with customer satisfaction. Traditional pricing strategies often depend on static rules or overly simplistic models that lack the ability to adapt to real-time changes in market demand, evolving customer behavior, and competitive trends. These outdated approaches can lead to missed revenue opportunities and suboptimal guest experiences. This research addresses these challenges by proposing a dynamic pricing strategy driven by deep reinforcement learning, which integrates real-time data streams with predictive analytics to create a highly responsive and intelligent pricing system. At the core of the proposed methodology is a novel framework that combines advanced deep neural network architectures with adaptive optimization algorithms. This framework is designed to optimize both pricing and inventory decisions across multiple booking channels simultaneously. The Innovative Pricing Transformer (IPT) model underpins this framework by leveraging attention mechanisms and temporal sequence modeling to accurately forecast future demand and recommend context-aware pricing decisions. In addition, the Adaptive Yield Optimization Strategy (AYOS) refines this process by incorporating real-world operational constraints such as overbooking policies, price parity requirements, and channel-specific pricing rules, ensuring practicality and compliance. Empirical analysis conducted on real-world datasets reveals that our approach consistently outperforms traditional pricing models, not only in revenue enhancement but also in improving overall customer satisfaction. The proposed strategy represents a scalable, efficient, and intelligent solution for modern hotel revenue management, enabling hotels to remain agile and competitive in dynamic and uncertain market conditions.

The network revenue management (RM) problem arises in airline, hotel, media, and other industries where the sale products use multiple resources. It can be formulated as a stochastic dynamic program, but the dynamic program is computationally intractable because of an exponentially large state space, and a number of heuristics have been proposed to approximate its value function. In this paper we show that the piecewise-linear approximation to the network RM dynamic program is tractable; specifically we show that the separation problem of the approximation can be solved as a relatively compact linear program. Moreover, the resulting compact formulation of the approximate dynamic program turns out to be exactly equivalent to the Lagrangian relaxation of the dynamic program, an earlier heuristic method proposed for the same problem. We perform a numerical comparison of solving the problem by generating separating cuts or as our compact linear program. We discuss extensions to versions of the network RM problem with overbooking as well as the difficulties of extending it to the choice model of network revenue RM.

No abstract available

No abstract available

No abstract available

No abstract available

Online Travel Platforms (OTPs) serve as crucial bridges between hotels and users, hotel staff can synchronize room inventory information with OTPs through manual and auto modes. In the manual mode, the hotel staff must manually maintain the inventory information on the OTPs. This mode often leads to the "inventory synchronization delay'' phenomenon where OTPs show no availability while hotels still have available rooms, seriously affecting the competitiveness of OTPs and hotel sales. To address this issue, Fliggy uses inventory add prediction (IAP) to determine whether to add an inventory for the sold-out room type. However, in practice, accurate modeling of IAP faces significant challenges due to the data sparsity. In this paper, we propose a Contrastive Learning framework for Inventory Add Prediction at Fliggy (CL4IAP), which consists of the Joint Pay-Accept Prediction Module, the Data Augmentation Module, and the Contrastive Learning Module. Specifically, the Joint Pay-Accept Prediction Module aims to predict the likelihood of generating an order and the hotel acceptance after adding an inventory. It also includes a specially designed correlation enhancement component that facilitates the expert prediction network's learning through knowledge transfer based on inter-task correlation. In the Data Augmentation Module, we design three novel data augmentation strategies for the first time based on the correlation and importance of features. In the Contrastive Learning Module, we design instance-level and cluster-level contrastive losses, which aim to minimize the distance between positive sample pairs and mitigate the negative impact of false negative sample pairs, respectively. Both offline and online experiments demonstrate the effectiveness of CL4IAP, and CL4IAP has been successfully deployed on Fliggy.

<jats:p>In this paper, we consider a two-commodity stochastic <jats:inline-formula> <math xmlns="http://www.w3.org/1998/Math/MathML" id="M1"> <mfenced open="(" close=")" separators="|"> <mrow> <mi>s</mi> <mo>,</mo> <mi>S</mi> </mrow> </mfenced> </math> </jats:inline-formula> perishable inventory system of the multiqueue Jackson network at a service facility with reneging and jockeying of the customers. The waiting room capacity of the first queue is <jats:inline-formula> <math xmlns="http://www.w3.org/1998/Math/MathML" id="M2"> <mi>M</mi> <mo>/</mo> <mi>M</mi> <mo>/</mo> <mn>1</mn> <mo>/</mo> <mi>∞</mi> </math> </jats:inline-formula> and the second queue is <jats:inline-formula> <math xmlns="http://www.w3.org/1998/Math/MathML" id="M3"> <mi>M</mi> <mo>/</mo> <mi>M</mi> <mo>/</mo> <mn>1</mn> <mo>/</mo> <msub> <mrow> <mi>C</mi> </mrow> <mrow> <mn>2</mn> </mrow> </msub> </math> </jats:inline-formula>, where <jats:inline-formula> <math xmlns="http://www.w3.org/1998/Math/MathML" id="M4"> <msub> <mrow> <mi>C</mi> </mrow> <mrow> <mn>2</mn> </mrow> </msub> <mo><</mo> <mi>∞</mi> </math> </jats:inline-formula>. In this network, different queues have different service rates. Service times are exponentially distributed. Customers join the system at a rate of <jats:inline-formula> <math xmlns="http://www.w3.org/1998/Math/MathML" id="M5"> <msub> <mrow> <mi>λ</mi> </mrow> <mrow> <mi>i</mi> </mrow> </msub> <mo>;</mo> <mfenced open="(" close=")" separators="|"> <mrow> <mi>i</mi> <mo>=</mo> <mn>1,2</mn> </mrow> </mfenced> </math> </jats:inline-formula> using a Poisson process. We assumed that some customers maybe impatient due to the long waiting time in the queue and can leave the system without service or may switch from the longest queue to the shortest queue to reduce the waiting time. If the inventory levels of the items from the warehouse <jats:inline-formula> <math xmlns="http://www.w3.org/1998/Math/MathML" id="M6"> <mi>i</mi> </math> </jats:inline-formula> at any time reaches below or reorder level <jats:inline-formula> <math xmlns="http://www.w3.org/1998/Math/MathML" id="M7"> <msub> <mrow> <mi>s</mi> </mrow> <mrow> <mi>i</mi> </mrow> </msub> </math> </jats:inline-formula>, an order <jats:inline-formula> <math xmlns="http://www.w3.org/1998/Math/MathML" id="M8"> <msub> <mrow> <mi>Q</mi> </mrow> <mrow> <mi>i</mi> </mrow> </msub> <mo>=</mo> <mfenced open="(" close=")" separators="|"> <mrow> <mfenced open="(" close=")" separators="|"> <mrow> <msub> <mrow> <mi>S</mi> </mrow> <mrow> <mi>i</mi> </mrow> </msub> <mo>−</mo> <msub> <mrow> <mi>s</mi> </mrow> <mrow> <mi>i</mi> </mrow> </msub> </mrow> </mfenced> <mo>></mo> <mn>0</mn> <mo>;</mo> <mi>i</mi> <mo>=</mo> <mn>1,2</mn> </mrow> </mfenced> </math> </jats:inline-formula> units is placed to bring the inventory level up to <jats:inline-formula> <math xmlns="http://www.w3.org/1998/Math/MathML" id="M9"> <msub> <mrow> <mi>S</mi> </mrow> <mrow> <mi>i</mi> </mrow> </msub> </math> </jats:inline-formula>. The joint probability distribution of the number of customers in the system and also the stock level of the warehouse is determined in the equilibrium case. Several important measures of steady-state system performance are derived. Several examples are presented to demonstrate the reneging and jockeying behavior of the customers for this system. A sensitivity analysis is performed to investigate the effect of changing the parameters on the total expected cost of the system.</jats:p>

Solutions to address the periodic review inventory control problem with nonstationary random demand, lost sales, and stochastic vendor lead times typically involve making strong assumptions on the dynamics for either approximation or simulation, and applying methods such as optimization, dynamic programming, or reinforcement learning. Therefore, it is important to analyze and evaluate any inventory control policy, in particular to see if there is room for improvement. We introduce the concept of an equilibrium policy, a desirable property of a policy that intuitively means that, in hindsight, changing only a small fraction of actions does not result in materially more reward. We provide a light-weight contextual bandit-based algorithm to evaluate and occasionally tweak policies, and show that this method achieves favorable guarantees, both theoretically and in empirical studies.

No abstract available

No abstract available

No abstract available

No abstract available

With the development of information technology, the hotel industry needs to utilize advanced technologies and methods to improve the accuracy and efficiency of room forecasting and pricing in order to adapt to the changes in the market and the pressure of competition. This paper adopts a combination of quantitative and qualitative methods, following two phases: (1) data analysis phase, using statistical analysis and visualization techniques to clean, process, describe, and explore the collected data related to hotel rooms; (2) model building phase, using machine learning techniques to construct and train a deep neural network model to achieve the prediction of hotel rooms, and using reinforcement learning techniques that realizes the pricing of hotel rooms. In this paper, theoretical and practical experiments are conducted to compare and evaluate with existing research methods and models from different perspectives and levels. The results show that the model proposed in this paper outperforms other models in all indicators, with higher prediction accuracy and pricing efficiency, as well as good generalization and adaptation capabilities. The research in this paper has important theoretical and practical significance for revenue management and competitive strategies in the hotel industry.

No abstract available

No abstract available

No abstract available

Time series forecasting aims to model the change in data points over time. It is applicable in many areas, such as energy consumption, solid waste generation, economic indicators (inflation, currency), global warming (heat, water level), and hotel sales forecasting. This paper focuses on hotel sales forecasting with machine learning and deep learning solutions. A simple forecast solution is to repeat the last observation (Naive method) or the average of the past observations (Average method). More sophisticated solutions have been developed over the years, such as machine learning methods that have linear (Linear Regression, ARIMA) and nonlinear (Polynomial Regression and Support Vector Regression) methods. Different kinds of neural networks are developed and used in time series forecasting problems, and two of the successful ones are Recurrent Neural Networks and N-BEATS. This paper presents a forecasting analysis of hotel sales from Türkiye and Cyprus. We showed that N-BEATS is a solid choice against LSTM, especially in long sequences. Moreover, N-BEATS has slightly better inference time results in long sequences, but LSTM is faster in short sequences.

Hotel Occupancy Rate is one of the important leading indicators for calculating the Accommodation Sub-Category of Gross Regional Domestic Product (GRDP). By the extreme decline of the Hotel Occupancy Rate data due to COVID-19 and the unavailability of current data to counting GRDP quarterly, the Hotel Occupancy Rate prediction needs to do with the appropriate forecasting method. The authors use data from Google Trends as an additional variable in predicting the Hotel Occupancy Rate using the ARIMAX model and then compares it with the ARIMA model. The results showed that the ARIMAX model had better accuracy than ARIMA, with a MAPE value of 9.64 percent and an RMSE of 4.21 percent. This research concluded that if there is no change in government policy related to social restrictions until the end of the year, the ARIMAX model predicts the December 2021 Hotel Occupancy Rate of 38.59 percent.

Hotel Room Sales prediction using previous booking data is a prominent research topic for the online travel agency (OTA) sector. Various approaches have been proposed to predict hotel room sales for different prediction horizons, such as yearly demand or daily number of reservations. An OTA website includes offers of many companies for the same hotel, and the position of the company’s offer in OTA website depends on the bid amount given for each click by the company. Therefore, the accurate prediction of the sales amount for a given bid is a crucial need in revenue and cost management for the companies in the sector. In this paper, we forecast the next day’s sales amount in order to provide an estimate of daily revenue generated per hotel. An important contribution of our study is to use an enriched dataset constructed by combining the most informative features proposed in various related studies for hotel sales prediction. Moreover, we enrich this dataset with a set of OTA specific features that possess information about the relative position of the company’s offers to that of its competitors in a travel metasearch engine website. We provide a real application on the hotel room sales data of a large OTA in Turkey. The comparative results show that enrichment of the input representation with the OTA-specific additional features increases the generalization ability of the prediction models, and tree-based boosting algorithms perform the best results on this task.

No abstract available

Purpose Mean absolute percentage error (MAPE) is the primary forecast evaluation metric in hospitality and tourism research; however its main shortcoming is that it is asymmetric. The asymmetry occurs due to over or under forecasts that introduce bias into forecast evaluation. This study aims to explore the nature of asymmetry and designs a new measure, one that reduces the asymmetric properties while maintaining MAPE’s scale-free and intuitive interpretation characteristics. Design/methodology/approach The study proposes and tests a new forecasting accuracy measure for hospitality revenue management (RM). A computer simulation is used to assess and demonstrate the problem of asymmetry when forecasting with MAPE, and the new measures’ (MSapeMER, that is, Mean of Selectively applied Absolute Percentage Error or Magnitude of Error Relative to the estimate) ability to reduce it. The MSapeMER’s effectiveness is empirically validated by using a large set of hotel forecasts. Findings The study demonstrates the ability of the MSapeMER to reduce the asymmetry bias generated by MAPE. Furthermore, this study demonstrates that MSapeMER is more effective than previous attempts to correct for asymmetry bias. The results show via simulation and empirical investigation that the error metric is more stable and less swayed by the presence of over and under forecasts. Research limitations/implications It is recommended that hospitality RM researchers and professionals adopt MSapeMER when using MAPE to evaluate forecasting performance. The MSapeMER removes the potential bias that MAPE invites due to its calculation and presence of over and under forecasts. Therefore, forecasting evaluations may be less affected by the presence of over and under forecasts and their ability to bias forecasting results. Practical implications Hospitality RM should adopt this measure when MAPE is used, to reduce biased decisions driven by the “asymmetry of MAPE.” Originality/value The MAPE error metric exhibits an asymmetry problem, and this paper proposes a more effective solution to reduce biased results with two major methodological contributions. It is first to systematically study the characteristics of MAPE’s asymmetry, while proposing and testing a measure that considerably reduces the amount of asymmetry. This is a critical contribution because MAPE is the primary forecasting metric in hospitality and tourism studies. The second methodological contribution is a procedure developed to “quantify” the asymmetry. The approach is demonstrated and allows future research to compare asymmetric characteristics among various accuracy measures.

The adoption of ICT in the hospitality industry is important in the development of business, minimizing costs and the generation of revenue as well as to reach more customers. Across the world, the use of technology and information and communication technologies (ICT) in the hospitality industry has significantly impacted the way hotels operate and the way they provide services to their guests. In the wake of turbulence in the hospitality industry especially as a result of the Covid-19 pandemic that reduced disposable income as well as crippling travel and tourism around the planet, hotels need to lower their operation costs, find ways of increasing their revenue as well as enhancing guest satisfaction. The study sought to establish the effect of technology and ICT policies on 4 and 5 star hotels in Nairobi City County. Specifically, the study sought to establish the effects of property management systems on guest satisfaction in 4 and 5 star hotels in Nairobi Kenya; to examine the effect of online booking platforms on guest satisfaction in 4 and 5 star hotels in Nairobi Kenya; to determine the effect of in-room technology on guest satisfaction in 4 and 5 star hotels in Nairobi Kenya and to establish the effect of customer relationship management systems on guest satisfaction in 4 and 5 star hotels in Nairobi Kenya. The study was guided by the Technology Acceptance Model (TAM) and the Resource Based View Theory. The target population was 4583 staff from 24 four and five star hotels in Nairobi. Stratified random sampling was used to choose a sample size of 368 employees. The data was collected using structured questionnaires for the employees. The questionnaires were administered through google online questionnaire. Regression analysis was conducted to establish the relationship between the variables. The results were presented in graphs and tables. The study used a multiple regression model to show the relationship between the study variables. The findings revealed an R squared coefficient of 0.643 and adjusted R squared of 0.618 at 95% significance level, implying that the technology and ICT policies adopted in the study (Property Management Systems, Online booking platforms, In-Room Technology, Customer Relationship Management Systems) jointly explained 64.3 percent of the variation in guest satisfaction in four and five star hotels in Nairobi City County. The study also found that property management systems had a positive and significant effect on guest satisfaction in four and five star hotels in Nairobi City County (β =.357, p=.013<.05); online booking platforms had a positive and significant effect on guest satisfaction in four and five star hotels in Nairobi City County (β =.427, p=.005<.05); in-room technology had a positive and significant effect on guest satisfaction in four and five star hotels in Nairobi City County (β =.322, p=.003>.05) and finally the study found that customer relationship management systems as an aspect of technology and ICT policies had a positive and significant effect on guest satisfaction in four and five star hotels in Nairobi City County(β =.383, p=.000<.05). The study concluded that technology and ICT policies had a positive and significant effect on guest satisfaction in four and five star hotels in Nairobi City County. The study thus recommended that four and five star hotels in Nairobi City County should consider adopting technology and ICT policies such as Property Management Systems, Online booking platforms, In-Room Technology, Customer Relationship Management Systems and others as ways of enhancing guest satisfaction. Keywords: Technology, Guest Satisfaction, ICT Policies

This systematic literature review (SLR) examines the adoption, challenges, and outcomes of emerging technologies on Hotel Information Systems (HIS) in the Indian hospitality sector. Synthesizing insights from 124 peer-reviewed sources from SCOPUS and ABDC-listed journals published between 2015 and 2025, this study focuses on technologies such as Artificial Intelligence (AI), Internet of Things (IoT), cloud computing, PMS, CRS, 5G, Big data etc. The review identifies key organizational and technological drivers impacting adoption, analyses prevalent barriers, and assesses resultant impacts on operational efficiency and guest satisfaction. The paper concludes with policy and managerial implications to foster broader technology diffusion and competitive advantage in Indian hotels.

The article examines the essence of the terms "price policy" and "concept of pricing policy". The relationship between the terms "pricing strategy", "concept of pricing policy" and "price policy" is shown. In the scientific literature, the term "pricing concept" is quite rare and is not fully researched. In some sources, pricing concepts are identified with pricing strategies, pricing policies, or pricing methods. In our opinion, the following definition of the term "concept of pricing" can be given, which, in turn, has several meanings: 1) it is a system of views on the essence of pricing processes; 2) fundamental ideas embedded in pricing theory, pricing methodology, pricing policy and pricing strategies; 3) a way of understanding pricing processes. We also offer the following definition of the concept of pricing policy – these are the basic ideas used in the pricing policy of the enterprise (company), regarding the main ways and methods of implementing the pricing strategy. The following classification criteria are proposed and relevant concepts of price policy in the hotel and restaurant business are highlighted: 1. In accordance with the transparency and comprehensibility of information about the cost of hotel and restaurant services (spontaneous market pricing policy; transparency of prices and conditions for providing hotel services; setting prices in the process of agreements with guests with the determination of the lower limit of the price of services). 2. In accordance with the planned level of profit from each visitor (maximizing the profits of hotels; availability of prices and services for as many visitors as possible). 3. Exclusivity of places and landscapes for recreation (exclusivity of hotel buildings and rooms; exclusivity of resorts and places around the hotel). 4. Stars of hotels. 5. Purpose of hotels 6. Seasonality 7. Innovativeness of service provision, etc. It was determined that the concepts of pricing policy are most often used in practice in combination with each other, in the form of combined concepts of pricing, which requires high qualification of management personnel in the hotel and restaurant business.

With the rapid advancement of the hotel industry, studies on boosting operational efficiency have largely focused on employee productivity and human-centric policies, yet there is a noticeable lack of strategies aimed at enhancing room occupancy rates. This paper address this gap by examining factors influencing hotel reservation cancellations and proposing effective mitigation strategies. Utilizing a decision tree methodology, the research identifies key elements that contribute to cancellations. The significance of this study lies in its contribution to optimizing room occupancy within hotel management. Practical strategies, such as lowering booking security deposits, are suggested to decrease cancellation rates. These insights enhance not only operational efficiency and resource utilization but also offer actionable advice for improving guest satisfaction. By embracing data-driven decision-making, the study promotes modernized management practices and supports sustainable growth in hospitality. Three aspects for improvement are highlighted: internal operations, customer-related factors, and external influences. Recommendations include reducing security deposits at booking to lower cancellations and optimize occupancy.

Inspired by interviews with hotel managers and analysis of real pricing data, we develop an experience-led pricing algorithm called the DDD algorithm, which derives actionable decisions from historical data. DDD stands for “Data Collation”, “Demand Learning,” and “Decision Optimization.” In data collation, data that are similar to the current scenario and valuable for learning are selected for demand learning. In demand learning, due to the scarcity of data, traditional demand learning methods are not effective. Therefore, we adopt a novel approach, the core idea of which is to infer the cause (i.e., demand parameters) from the outcome (i.e., decisions). In decision optimization, the demand parameters are used to determine the final pricing decision, with a balance between exploration and exploitation. We evaluate the DDD algorithm based on revenue regret and computational complexity. The revenue regret of the DDD algorithm is O(TK), and its time complexity is O(N3+KL5.5), where N is the size of the historical dataset, T is the decision times, K is the quantity of selected data and L represents the number of days in advance for sale. Finally, the effectiveness of the DDD algorithm is demonstrated through numerical experiments using real hotel data.

Accurately capturing the behavioral factors of different types of customer groups and adopting targeted service strategies is the key to business competition in the hotel industry. In this paper, we combine the variance Boston matrix and PSO-based K-means algorithm to achieve hotel customer attribute segmentation based on customer behavior, customer value and word-of-mouth reliability, and then use deep learning algorithms to construct a hotel customer behavior prediction model. The feature fusion layer and SENet are incorporated into the residual network in order to utilize the feature expression ability of different layers and the spatial coding ability between different channels to enhance the hotel customer behavior predictive ability. Downloading the public dataset from the online wine travel platform for example analysis, it is found that the classification of this paper's algorithm before customer segmentation has a correct rate of 83.75%, which is higher than the rest of the baseline models. After customer segmentation this paper's algorithm achieves the highest recall rate in all customer categories, and the recall rate is as high as 84% on category 1 customer groups, and the superiority of the designed algorithm is verified. This study facilitates hotel management to target customer service and retention according to different customer groups.

The hotel industry is highly competitive and faces challenges, such as fluctuating demand, intense competition, and shifting consumer preferences. One critical issue that hotels frequently encounter is the cancellation of room reservations, which disrupts operational planning and resource management and leads to significant financial losses. Accurately predicting the likelihood of reservation cancellation is essential to mitigate these negative impacts and optimize revenue management strategies. This study focuses on the development of a predictive model for hotel room reservation cancellations using a decision-tree algorithm. The Decision Tree was selected for its ability to manage complex relationships between variables and ease of interpretation, making it accessible to hotel managers without technical expertise. To enhance the performance of the model, a forward selection technique was employed to identify the most relevant features, ensuring a balance between the model complexity and predictive accuracy. Additionally, resampling techniques were applied to address class imbalance in the dataset, which is common in cancellation cases where non-cancelled reservations outnumber cancelled reservations. This study explores the prediction of hotel room reservation cancellations using a decision tree algorithm enhanced by feature selection and resampling. The model achieved an accuracy improvement to 90%, with precision and recall each increasing by 5,5% after applying these techniques. These findings suggest practical applications for improving cancellation predictions and optimizing revenue management strategies for hotels. The study provides insights into how data-driven approaches can enhance decision-making processes within the competitive hospitality industry.

The significance of having proper revenue management and convenient operations in hotels rely on accurate daily demand. The main challenges that are being tackled conventionally are the forecasting of the number of cancellations and the calculation of the Average Daily Rate (ADR). One of the distinct characteristics of the hotel industry being so unstable is to adopt proactive strategies by dealing with a lot of external changes such as pandemics, disasters caused by nature, and economic fluctuations around the world. Booking cancellations are instrumental in helping hotel managers to optimise resources and inventory while ADR forecasting offers them equally essential projections concerning anticipated profit or loss margins. This research relies on several data modelling steps, including time series aggregation and decision merging, which are later followed by decomposition and model selection. SARIMAX and LSTM models are adopted for future traffic flow modelling, which demonstrates better forecasting performances. Binary classification is employed for feature engineering techniques together with model selection methods. Binary Classification is performed with a number of experiments with machine learning algorithms, AdaBoost turned out to be the best model which surpassed CART , KNN , Random Forest, Gradient boosting algorithms , and Light Gradient Boosting algorithms. The results of this are of great help to the hotel management for taking better decisions which are connected to the changing situations of the market

No abstract available

ABSTRACT The Hotel Demand Forecasting Model leverages the ARIMA model to enhance safety and operational efficiency in the hospitality industry. By integrating historical data, market trends, and safety parameters, the model accurately forecasts guest demand amidst unpredictable factors like fluctuating travel trends and evolving safety regulations. Traditional models often fail to account for these variables, leading to inefficiencies and safety concerns. This model aims to establish a robust system that considers the interconnectedness of Property Management System (PMS) data, staff productivity, and procurement efficiency, optimizing hotel operations. The research employs Exploratory Data Analysis (EDA) with Multivariate Graphical techniques to uncover patterns and trends impacting demand. Implemented in MATLAB, the proposed model achieves 91% accuracy, offering a strategy that balances automation, efficiency, and safety. By automating demand prediction, hotels can optimize resource allocation, proactively respond to demand fluctuations, and prioritize guest satisfaction while maintaining safety protocols.

This research investigates the effectiveness of machine learning (ML) for predicting hotel cancellations, providing valuable insights for optimizing hotel management strategies. ML algorithms, including LightGBM, Gradient Boosting Classifier, AdaBoost Classifier, and CatBoost Classifier, consistently achieve exceptional accuracy of 99%, making them robust choices for cancellation prediction. While XGBoost demonstrates commendable performance with a 98% accuracy rate, the study underscores the importance of precision and recall metrics for effective decision-making in the hospitality sector. These findings guide decision-makers in the hotel industry towards informed choices in ML algorithms, emphasizing the nuanced trade-offs between precision, recall, and computational efficiency. The research significantly contributes to advancing predictive modelling in the hospitality sector, laying the groundwork for enhanced reservation systems and customer service practices in the hotel industry.

This paper aims to predict whether guests will cancel hotel reservations by analyzing hotel operation data and using deep learning algorithms. The cancellation of hotel reservations is a common problem in the world. When the hotel keeps the guest room and the customer cancels the reservation, it will bring difficulties in management and operation as well as economic losses to the hotel. With advances in data collection and storage technologies, we are faced with massive amounts of data that are difficult to analyze. Deep learning has the ability to learn from massive amounts of data and make effective predictions. To address this issue, our deep learning model examines and forecasts extensive data related to hotel reservations. We employ a Long Short-Term Memory neural network (LSTM) optimized through the Chaos Sparrow Search Algorithm (CSSA). This approach enables us not only to comprehend the dynamics of the cancellation rate but also to pinpoint potential reservation cancellations by identifying specific customer patterns.

The forecasting of demand or cancellations is highly important for efficient revenue management in the hotel industry. Previous studies have mainly focused on the accuracy of the prediction of reservation number or cancellation rate on a specific accommodation or hotel chain; therefore, the application of the prediction to different accommodations or under the behavioral change of customers in response to natural or human events is difficult without the re-estimation of the prediction model. Information of the customer behavioral trend on the accommodation reservations is necessary for the construction of a general forecasting model. In this study, we focus on one of the general trends of customer behavior, that is, the reservation timing and the time changes of the cancellation probability using the big data of the reservation records provided by an online trip agency in Japan. We showed that the reservation timing and cancellation probability can be decomposed by five and six exponential functions of the days until the stay and the days from the reservations. We also showed that the significant factors influencing the time changing patterns are the guest numbers per room for both reservation and cancellation, composition of guests in terms of the number and gender of guests, and the stay length for reservation. These findings imply that the customer behavior during accommodation reservation could be categorized into multiple motivational factors toward reservations or cancellations. Our results contribute to the construction of a general forecasting model on the accommodation reservations.

Hotel booking cancellations significantly affect revenue management. While existing studies have explored various machine learning models, a systematic analysis of the interplay between feature selection, missing value imputation, and model choice is lacking. This study bridges this gap by conducting a comprehensive evaluation of three imputation methods (mean, kNN, deletion) across three model architectures (Logistic Regression, DNN, LSTM), utilizing features selected by XGBoost importance. Our results on a real-world dataset show that: (1) using only the top three features (lead time, booking changes, total of special requests) achieved comparable performance to using all features, promoting model efficiency; (2) the optimal imputation strategy is model-dependent; kNN imputation worked best for DNN, while direct deletion yielded the best generalization for LSTM; and (3) incorporating dropout effectively mitigated overfitting in LSTMs, improving F1-score by nearly 3 percentage points. This work provides data-driven guidelines for building robust cancellation prediction systems in the hospitality industry. More broadly, our findings highlight that the optimal data preprocessing pipeline is not one-size-fits-all but is intrinsically tied to the inductive biases of the model architecture, a principle likely applicable to other sequential decision-making domains.

The rapid changes in the global economic situation have brought many challenges to the hotel industry. Many hotels face operating difficulties due to the lack of a stable source of income. Customer cancellation is one of the main factors leading to revenue instability, however, a large proportion of hotels are still not effectively taking measures to deal with it. The paper uses machine learning methods to fix profitability issues in the hotel industry. First, the paper processes the raw data from the reservation conditions of a hotel. Then, some key variables are selected through a correlation matrix. The paper introduces and explains some machine learning models to show the advantages and disadvantages of each model. After programming, compare the accuracy scores of models. In the last, select the best model. In the paper, logistic regression, single decision tree, random forest, neural net, and boosted tree are alternative options. The random forest model has the highest accuracy score. The reasons for the conclusion and the suggestions for hotels are listed in the section of the discussion.

Hotel reservation cancellations pose significant operational and financial challenges for the hospitality industry. With the growing prevalence of online booking platforms and flexible cancellation policies, accurately predicting whether a reservation will be canceled has become increasingly critical for revenue management and resource optimization. This study investigates and compares a range of machine learning and deep learning models -including XGBoost, Random Forest, TabNet, PyTorch-based neural networks, and Logistic Regression- for their effectiveness in predicting booking cancellations using a publicly available dataset comprising 36,275 reservations. Each model was evaluated using 5-fold stratified cross-validation, with performance assessed via accuracy, F1 score, and area under the ROC curve (AUC). Ensemble methods (XGBoost and Random Forest) achieved the best predictive performance (AUC scores of 0.9526 and 0.9553, respectively), outperforming both traditional statistical models and deep learning alternatives. Analysis revealed that variables such as lead time, number of special requests, and market segment type are consistently strong predictors of cancellation behavior. The results highlight the potential of interpretable machine learning models to support proactive decision-making in hotel operations. By integrating these models into reservation systems, hotels can reduce revenue loss, better manage capacity, and personalize customer engagement strategies. This research offers a robust benchmarking framework and practical insights for applying predictive analytics in the hospitality domain.

With the rise of online travel platforms (e.g., Ctrip, Qunar), modifying or canceling hotel reservations has become more convenient, which in turn increases the probability of reservation cancellations. Each cancellation imposes issues on hotels, such as higher room vacancy rates, unstable revenue, and increased operating costs, presenting a significant challenge to revenue management teams. Only by accurately predicting cancellations can effective strategies be formulated to mitigate such losses. To address this problem, this paper enriches existing hotel reservation information by applying feature interaction technology. Subsequently, we compare the performance of machine learning models, including Decision Tree (DT), Random Forest (RF), and Gradient Boosting (GB), and selects the optimal model based on evaluation results. Through the methods, hotels can proactively identify the likelihood of reservation cancellations, thereby developing more targeted strategies to reduce potential losses.

Hotel reservations have become a prevalent choice for customers. However, cancellations of these reservations present a significant challenge for hotels, potentially resulting in financial losses and a decline in customer satisfaction. To address the issue of improper management of cancellations and minimize losses, machine learning can be employed to analyze and predict cancellations based on customer information. In cooperative scenarios where hotels collaborate to train a unified model, traditional algorithms that aggregate all data raise concerns about the protection of sensitive customer information. In this context, federated learning emerges as an effective solution to ensure the privacy protection of customers while achieving the desired predictive outcomes. Thus, to protect customer privacy while preserving performance of the federated learning model in comparison to the non-federated version, this work proposed to implement both federated learning and non-federated algorithms based on neural network to make predictions of cancellation based on multiple factors. The federated learning approach achieved a final testing accuracy of 76.64%. Although this accuracy was about 9% lower than the non-federated case, its loss was over half the loss of non-federated, and its testing accuracy was similar to the training accuracy, while the non-federated algorithms testing accuracy was approximately 4% lower than the training one. Such results indicate that although accuracy was relatively lower, the federated learning approach prevented the overfitting problem in the non-federated case, while the data privacy problem was resolved.

This study applies Bayesian models to predict hotel booking cancellations, a key challenge affecting resource allocation, revenue, and customer satisfaction in the hospitality industry. Using a Kaggle dataset with 36,285 observations and 17 features, Bayesian Logistic Regression and Beta-Binomial models were implemented. The logistic model, applied to 12 features and 5,000 randomly selected observations, outperformed the Beta-Binomial model in predictive accuracy. Key predictors included the number of adults, children, stay duration, lead time, car parking space, room type, and special requests. Model evaluation using Leave-One-Out Cross-Validation (LOO-CV) confirmed strong alignment between observed and predicted outcomes, demonstrating the model's robustness. Special requests and parking availability were found to be the strongest predictors of cancellation. This Bayesian approach provides a valuable tool for improving booking management and operational efficiency in the hotel industry.

The global tourism industry is expanding rapidly, making effective management of hotel booking cancellations crucial for improving service and efficiency. Existing models, based on static data assumptions and fixed parameters, fail to capture dynamic changes and temporal trends. Real-world cancellation decisions are influenced by factors such as seasonal variations, market demand fluctuations, holidays, and special events, which cause significant changes in cancellation rates. Traditional models struggle to adjust dynamically to these changes. This article proposes a novel approach using deep reinforcement learning techniques for predicting hotel booking cancellations over time. We introduce a framework that combines dynamic temporal reinforcement learning with policy-enhanced LSTM, capturing temporal dynamics and leveraging multi-source information to improve prediction accuracy and stability. Our results show that the proposed model significantly outperforms traditional methods, achieving over 95.9% prediction accuracy, a model stability of 0.98, an F1 Score approaching 1, and a mutual information score of approximately 0.93. These results validate the model’s effectiveness and generalization across diverse data sources. This study provides an innovative and efficient solution for managing hotel booking cancellations, demonstrating the potential of deep reinforcement learning in handling complex prediction tasks.

No abstract available

This study focuses on predicting hotel booking cancellations using machine learning to improve accuracy and operational efficiency. The methods used include Random Forest, Artificial Bee Colony (ABC), and Gradient Boosting Decision Tree (GBDT). ABC, which excels in optimization but is prone to local optima, is combined with GBDT for feature selection. The dataset used is Hotel_Bookings from Kaggle, with 119,390 entries and 28 features. The data is processed through cleansing, normalization, and split into 75% for training and 25% for testing. Feature selection is performed using ABC and GBDT, and the prediction model is built using Random Forest. Model evaluation using confusion matrix and metrics like precision, recall, f1-score, and accuracy shows accuracies of 86.17% and 86.65% for ABC and GBDT, respectively. Increasing the number of trees and features generally improves model performance, with feature selection showing significant performance improvements compared to models without feature selection.

Abstract: Hotel managers find it beneficial to predict hotel booking cancellations as it enables them to enhance room inventory management, pricing strategies, and customer satisfaction by proactively addressing potential problems. In this study, we present a machine learning-centered method for forecasting hotel booking cancellations. Overall, our proposed approach will equip hotel managers with a robust tool for predicting hotel booking cancellations and a deeper understanding of the factors influencing them. This, in turn, will enable data- driven decision making to optimize room inventory, pricing strategies, enhance customer satisfaction, and ultimately boost revenue

No abstract available

Online travel sales continue to increase every year. Recorded in 2019, digital transactions related to online travel reached USD 755.4 billion. One of the supports of the travel business is the tourism and hospitality industry. The online reservation system is one of the most attractive solutions in the hospitality industry. Cancellation of hotel bookings or reservations through the online system is currently one of the problems in the hotel management system. When the reservation has been canceled, the hotel will be harmed. Therefore, predicting whether a booking will be canceled or not using the help of data science is needed so that the hotel can minimize lost profits. Therefore, by using datasets related to hotel booking requests, a predictive analysis using the CRISP-DM framework is conducted. By first performing some data preparation processes, this study uses a tree-based algorithm to perform the prediction. The experiment produced that Random Forest model has the best value with an accuracy value of 0.8725 and it is found that the time difference between booking is made and arrival time is the most influential feature in predicting the level of cancellation.

Enhancing the accuracy of short-term forecasts for cancellation rates offers revenue managers the opportunity to formulate a pricing strategy for the upcoming day, yielding favorable economic outcomes. This study proposes a methodology based on dynamic models utilizing machine learning methods such as LightGBM, XGBoost, Random Forest, and ANN-MLP, highlighting the importance of data dimensionality reduction while using higher-performing variables to improve predictive accuracy and stability. Notably, lagged variables within a few days of the forecasted date and reservations made through OTAs and BAR-rated reservations exhibit significant predictive power. The results indicate that artificial neural network multi-layer perceptron (ANN-MLP) outperforms other models, especially in longer forecast horizons. The study recommends adaptable strategies considering historical data and temporal trends and leveraging ANN-MLP for superior accuracy. The findings offer valuable insights for industry practitioners, providing a nuanced understanding of cancellation patterns and suggesting strategies to optimize cancellation prediction models in a competitive marketplace.

Abstract - Hotel booking cancellations create significant challenges for the hospitality industry by affecting revenue management, demand forecasting, and resource allocation. Traditional prediction methods often fail to capture the complex behavioral patterns associated with customer cancellations. This research proposes a machine learning-based approach using a Multi-Layer Perceptron (MLP) classifier to predict hotel booking cancellations accurately. The model analyzes various booking attributes such as lead time, room type, customer segment, location, and booking history to detect patterns associated with cancellation behavior. Data preprocessing techniques including normalization, handling missing values, and encoding categorical variables are applied to improve model performance. Hyperparameter tuning is performed to optimize hidden layers, activation functions, and learning rates. Experimental results demonstrate that the proposed model provides high prediction accuracy and supports hotel managers in making informed decisions related to demand forecasting, overbooking strategies, and revenue optimization. The system ultimately contributes to improved operational efficiency and customer satisfaction in the hospitality sector. Hotel booking cancellations create major challenges for the hospitality industry by affecting revenue management and demand forecasting. This research proposes a machine learning approach using a Multi-Layer Perceptron (MLP) classifier to predict hotel booking cancellations based on historical booking data. The model helps hotel managers improve decision-making, optimize resources, and reduce financial losses caused by cancellations. Key Words: Machine Learning, Hotel Booking Cancellation, Multi-Layer Perceptron, Hospitality Management, Predictive Analytics.

. The online hotel reservation channels have caused increased cancellations, which is a revenue-diminishing factor for the hotels to deal with. Therefore, it is important to understand the bookers’ behavior and make good prediction on the cancellation decision. In this work, we analyzed the characteristics of distinct sub-populations among the bookers by a hotel reservation data. We built a range of machine learning models to make predictions on whether a customer is going to cancel the reservation. We also improved our methods by selecting the important features in the data. More-over, we investigated deep into the important features to find rational explanation on the effect. Hopefully, our work can provide suggestions on room management for hotels.

In recent years, machine learning has emerged as a powerful tool with widespread applications across various domains due to its ability to process and analyze vast amounts of data. This study explores the application of machine learning techniques in predicting hotel booking cancellations using Property Management System (PMS) data. The research involves a comprehensive process, including data cleaning, feature engineering, feature selection, and model development. Feature selection and dimensionality reduction using Principal Component Analysis (PCA) and Lasso regression identified key predictive features, facilitating the rapid creation of neural network models. A diverse set of machine learning and deep learning models, such as Logistic Regression, Decision Tree, Random Forest, XGBoost, Multilayer Perceptron (MLP), Convolutional Neural Network (CNN), Deep Neural Network (DNN), and Long Short-Term Memory (LSTM), were employed. All models achieved accuracies exceeding 80%, with neural networks nearing 100%. These results highlight the efficacy of these models in predicting cancellations across different hotels, revealing consistent cancellation patterns. The study demonstrates the potential of machine learning to optimize hotel management by accurately forecasting booking cancellations, thereby reducing uncertainty and increasing revenue. Future work may focus on exploring more advanced feature engineering techniques and models to further enhance prediction accuracy and generalizability.

Hotel booking cancellation prediction is crucial in conducting revenue and resource management for hotels. This paper provides three possible substitutes for the neural network including logistic regression, k -Nearest Neighbor ( k - NN), and CatBoost, whereas CatBoost, is the most suitable model for hotels to do the prediction. The advantages of them are effectiveness, high accuracy, and lower cost. The dataset used in this paper was adapted from Kaggle, a set of the booking data from two types of hotels (resort hotel and city hotel) in Portugal, and the corresponding customers’ information. We select some key variables as the predictor to train and test the prediction models based on three machine learning algorithms. After preprocessing the raw data, i.e., standardizing, dealing with missing data, recoding some variables, and scaling, we conduct the prediction and compare each model through three metrics (confusion matrix, accuracy score, and 1 F -score). The result indicates that CatBoost has the best performance in predicting hotel booking cancellation because it has the greatest number of correct prediction samples and the highest accuracy score. We focus on the efficiency and economy of doing cancellation prediction in the hospitality industry to form a basis for future revenue and resource management for hotels.

There are currently a substantial amount of hotel reservations that are canceled owing to customer absence or cancellation. They may cause a great deal of inconvenience for the hotel, impair its efficiency or revenue, etc. Through examining a sample dataset about hotel reservations from Kaggle, the purpose of this research is to identify some basic information and features in this dataset, then describe six machine learning models, including KNN, Random Forest (RF), Decision Tree (DT), Logical Regression (LR), SVM and neural network using the Python programming language, and train them on this dataset. The next step is to compare results to one another. In order to get efficient booking outcomes, it is necessary to select the most effective method, which is Random Forest (based on their value of accuracy), to predict the future state of hotel reservations, i.e. whether the consumer will confirm or cancel the reservation. This study seeks to assist novices in gaining a deeper understanding of large data, the principles of some machine learning models, and the capacity to predict data.

No abstract available