Improving Time-Series Demand Modeling in Hospitality Business by Analytics of Public Event Datasets

Abstract

Forecasting occupancy in hospitality business with autoregressive time-series models does not intercept occasional impact of public events. Our goal was to find appropriate datasets and enrich existing predictive models to account for rare and explicable demand surges. The paper proposes processing framework: data source types and formats, and forecast algorithms based on natural language processing. The study shows that classical models using word collocations outperform state of the art deep neural networks. Also, the collocations that turn out to be important, occupy certain locations in a graph that represents the natural language. The findings may result in yet improved forecasts, leading to smarter offer pricing and, finally, increased competitiveness in hospitality business. They may also serve public interest in areas like parking management or public transport planning. INDEX TERMS Time series analysis, recurrent neural networks, regression analysis, natural language processing, predictive models.

Daily hotel demand forecasting with spatiotemporal features

“…Apart from developing new models, integrating other external factors that affect hotel demand into the model is also an effective way to improve forecasting accuracy. Existing external factors often considered in forecasting models include quarterly gross domestic product (GDP) (O’Neill and Ouyang, 2020), income levels (Song et al , 2011), location (Zheng et al , 2020), weather information (Pan and Yang, 2017), website traffic (Yang et al , 2014) and rare events (Kamola and Arabas, 2020). With the growth of online booking websites, advanced booking data play an important role in improving the accuracy of forecasting.…”

“…Moreover, short-term demand data shows stronger volatility and nonlinearity, adding difficulty for models to generate accurate forecasting results (Wu et al , 2017). Most existing hotel demand forecasting methods aim to capture the trend and seasonality of time series (Kamola and Arabas, 2020; Tang et al , 2015), but their forecasting performance declines significantly in the face of data instability (Yang et al , 2014). Therefore, recent studies promote the development of this field in two ways: model innovation and the introduction of explanatory variables.…”

Hotel demand forecasting: a comprehensive literature review

“…Some future reservation data may be known in advance when forecasting reservation demand (Tse and Poon, 2015), so how to handle the known reservation data is a key problem in reservation demand forecasting. Existing studies mainly solved this problem on the basis of two basic assumptions: known and future reservations are independent of each other, and a proportional relationship exists between known and future reservations (Kamola and Arabas, 2020). To obtain forecasts, a solution to the first hypothesis can be obtained by adding the average past reservations to the existing booking demand.…”

“…Cancellation data also reflect changes in hotel demand in the future; thus, ignoring this important information when forecasting hotel demand may lead to significant prediction errors (Sánchez et al , 2020). Existing studies typically used two methods to manage cancellation data (Kamola and Arabas, 2020). The first method involves using cancellation data as a time series to predict the cancellation of hotel orders in the future (Sánchez-Medina and C-Sánchez, 2020), and the second method involves applying cancellation data to classify cancellation orders then predicting the cancellation probability of a reserved order (Antonio et al , 2017; Sánchez et al , 2020).…”

“…Multisource data refer to the combination of historical data with other relevant factors and are commonly used in hotel demand forecasting. Factors considered in hotel demand forecasting models include economic factors, such as hotel prices (Tsang and Benoit, 2019) and tourist income (Song et al , 2011); and environmental factors, such as weather condition (Pan and Yang, 2017), destination website flow (Yang et al , 2014), competing hotel demand (Assaf and Tsionas, 2019) and special events (Kamola and Arabas, 2020). Prediction models based on such combined data aim to generate hotel demand predictions by revealing the relationship between hotel demand and the explanatory variables (Peng et al , 2014).…”