Statistical Empirical Analysis Of Hotel Comment Emotion

With the rapid development of the internet,more and more goods are sold online,not only include physical commodities,but also include many service commodities(hotel reservation.?Online taxi.?Homemaking service……).Consumers are not only the acquirers of commodity information,but also the exporters of information.They can evaluate,describe and recommend commodities.Massive commodities and massive data also lead to the exponential growth of Internet information.Through the analysis of these textual information,we can identify the potentially valuable content and understand the emotional tendency of these information-positive or negative,which will help to dig out the more valuable content,help businessmen to obtain new market opportunities,give consumers more and better consumer opinions and bring convenience to people's lives.However,with such a large amount of information and data,it is a huge workload to interpret the information only by human resources in order to understand the user's evaluation emotions of commodities,and there are still many subjective awareness and objective environment effects.Therefore,in order to more conveniently and quickly understand the user's emotional trend of commodity evaluation and extract effective information from large data,this paper adopts machine learning and deep learning methods to begin emotional classification analysis by 10000 hotels reviews from Teacher Tan's data set and 5000 hotels review data sets obtained by web crawlers.These can give consumers and businesses a more intuitive understanding of commodity information.Finally,through a series of model training,this paper chooses an ideal model,which can predict the emotional tendency of the commentary text data,and more directly analyze the positive or negative emotions of consumers.