Using an additive model to forecast seasonal fluctuations in the hospitality sector

Abstract

The study is devoted to the modeling and forecasting seasonal fluctuations in the hospitality industry to align them in a timely manner. It will increase the level of operation of fixed assets, employment, rational use of key groups of resources and contribute to the development of measures to reduce seasonal irregularity in the studied sphere, identify socio-economic consequences of seasonality not only at the level of an individual organization, complex, region, but also the country as a whole. The model with additive component (variation of variable values in the form of trend, seasonal and random components), taking into account the data on the number of tour packages sold by travel agencies for the last fourteen quarters, is used as a basic one. The algorithm of seasonality analysis is represented by three basic steps, including calculation of seasonal component values, decrease in the seasonality factor of data and trend, and assessment of forecast accuracy. To eliminate the influence of the seasonal component, the moving average method is used, which is based on the replacement of the actual levels of the dynamic series by the estimated ones with a much smaller variation (fluctuation). In the process of the study we built models, in particular, linear, power, exponential and logarithmic ones, and selected a trend model (power model in this case), which provides the highest approximation accuracy. The constructed model allowed to determine the forecasted number of sold tour packages, taking into account seasonal and trend components. It is concluded that the study of seasonal fluctuations in hospitality on the basis of the built model will allow to develop targeted measures to eliminate or smooth the seasonality, develop sustainable demand and increase the availability of the product (service) of hospitality.