Research on emotion analysis of Ctrip hotel reviews and user demand identification based on big data

Abstract

With the rise of e-commerce economy, more and more people choose to book hotel travel online. As Ctrip is in a leading position in the e-commerce travel platform industry, this paper takes Ctrip hotel review as the representative for research. First, this article selects the top 40 hotels in the Beijing recommendation page to collect user reviews and related comments. A total of 34,988 comments were obtained after data consolidation. Then, the text data is cleaned, word segmentation, stop word and other preprocessing work. Finally, the processed data were classified. The emotion classification adopted in this paper mainly includes the emotion dictionary and the machine learning algorithms. The classification of the emotion dictionary is used to define the general classification of the text data and to visualize the display, then the word vector is constructed and the data set is divided using the TF-IDF algorithm. According to the corresponding classification algorithm to train and test the training set and test set respectively, the classifier with the best emotion classification is the logistic regression model. It can be seen from the classification results of emotional polarity that the overall emotional tendency of users to Beijing hotel experience is positive. It is best to visually analyze the conclusions and results obtained, which not only has certain reference value for the hotel merchants and other subjects, but also provides theoretical reference for users.

In addition, this paper proposes a user demand identification and evolution analysis model based on online comment mining based on the data of Beijing hotel reviews on Ctrip APP. Using the Kano model and the LDA model, the review underwent classification, identification, feature emotion pair analysis as well as time series analysis. The results show that according to the emotional trend prediction, the emotional value of type 1, type 2 and type 3 of users showed an upward trend, while the emotional value of type 4 showed a downward trend. Users mainly focus on hotel services and environmental experience. The research has improved the time dimension analysis method and model of online reviews, and provided a reference value for analyzing user needs and predicting the emotional trend of hotel selection.