Application of Content Based Recommendation System in Homestay

In order to overcome the problems of the application of recommender system in homestay industry is not perfect and can not provide personalized choice for users, a recommendation system for homestay is proposed. The system is based on content recommendation. First of all, from a large number of users’ comments on homestay, we use Jieba and stop word list to make a Chinese word segmentation corpus. Then we use word2vec training corpus to get homestay label features. Then the TF-IDF model is used to vectorize the tag features, and the similarity value of the feature vector is calculated by cosine similarity to get the push value of the new user, and the user preference table is constructed by selecting tags. When the user completes the subscription, the user preferences table is updated to complete the recommendation. The system designed in this paper show that the accuracy and real-time of recommendation, improves user satisfaction, and provides personalized recommendation for users.