Research On Personalized Recommendation Of Hotel Based On Crowd Geographic Data

Due to the development of Internet and the wide use of smart phones,the rise of e-commerce websites has also been promoted.In particular,various kinds of tourism mobile phone software(Ctrip,meituan,etc.)are used by more and more users.It can be said that network tools have become an important tool for people to choose their accommodation.At present,the hotel recommendations provided by domestic mainstream tourism websites are popular and universal.In this paper,the hotel information data and POI data are used as the data sources,and the improved TF-IDF feature extraction algorithm and the hotel recommendation algorithm based on multi-dimensional feature clustering and user rating are used Vue.js Technology as a technical support,the design and development of personalized recommendation system,to achieve the hotel recommendation algorithm.In the process of data acquisition and processing.We choose Qingdao City as the destination,take Ctrip as the data source website,use Python web crawler technology to crawl the basic information of the hotel,user comment data,and POI data,and make a visual analysis of the hotel data and POI data.In the process of hotel feature tag extraction,based on the traditional TF-IDF algorithm,the discrete DI is added,and the TF-IDF feature extraction algorithm is improved.Compared with the traditional TF and TF-IDF algorithm,the accuracy and recall rate of the algorithm are improved.The algorithm is used to extract Hotel feature words and tags,and based on K-means clustering and hotel professional knowledge The clustering of hotel feature tags is done.In the process of personalized hotel recommendation.According to the travel mode,travel destination,and the extracted Hotel feature tag,the user can cluster the multi-dimensional features,and combine the user score to recommend the hotel.The first n-digit output of the recommendation result is given.Finally,the MAE value of the recommended algorithm is compared and verified.The experimental results show that it has good stability.In the implementation of hotel personalized recommendation system.The system realizes the selection according to the user's travel mode,travel characteristics and travel destination,carries out the personalized recommendation of the hotel combined with the user's score,and realizes the visualization of the recommendation results.Through the algorithm comparison and verification,the method proposed in this paper can accurately and effectively recommend hotels,which has certain research significance.