Analysis Of Wuhan Hotel Pricing Model Based On Machine Learning

With the rapid development of China's tourism industry,the development prospect of hotel industry is very optimistic.For many investors who invest in the hotel industry,hotel pricing is an important means to maximize their profits.A reasonable pricing strategy will also bring competitiveness to businesses.At the same time,the information of online hotels is open and transparent,and reasonable pricing strategies are formulated,which also brings greater satisfaction to consumers.Therefore,the analysis of hotel pricing has a wide range of application significance.The research of hotel pricing mainly focuses on the influencing factors of hotel price and hotel pricing strategy.The hotel pricing methods are mainly cost oriented,market-oriented and the combination of them.The cost guidance method ignores the influence of market on pricing.The two methods are difficult to obtain.The pricing models in the market orientation method take into account the influence of market supply and demand on Hotel pricing.The pricing method based on Internet is also considered in the pricing mode of Internet plus hotel marketing mode.Therefore,this paper studies the hotel pricing model of Wuhan City Based on the competition centered method in the network pricing method.The main contents are as follows:In this paper,8968 samples and 52 characteristic variables,including hotel price,hotel facilities,traffic environment,hotel policy and other features variables,are obtained by using reptile technology.First of all,the data preprocessing and feature selection are carried out for the original data,and the data after preprocessing is described and analyzed to get the research significance between features and prices.Secondly,use the data set after feature selection to learn the following models respectively:LASSO regression model,the K nearest neighbor regression model,the XGBOOST regression model,the Random Forest regression model and the model of the forest.And,the cross validation and grid search method are used to find the optimal parameters of the model and establish the optimal single model.Then,the first mock exam is integrated into the optimal single model by using the improved Stacking algorithm XGBOOST-CStacking model.The training features of the model are selected as the 86%important features of the XGBOOST model.The secondary learning device is built up by K nearest neighbor,Random Forest and XGBOOST to establish an improved Stacking model.The mean square error of the model test set is 0.0067,0.0052 and 0.0056.Therefore,the sub learner of the optimal XGBOOST-CStacking model is random forest.Finally,the MSE of test set is used to compare the five models.From the mean square error of the model,it can be seen that the improved Stacking model has the best effect in Wuhan Hotel pricing strategy,and its predicted mean square error is the smallest,which is 0.0052.