A Deep Learning Recommendation Algorithm Incorporating Topic Scoring For Tourist Attractions

As people's living standards continue to rise,so does their demand for tourism,and recommending attractions with different characteristics to different users is gradually becoming an important issue in the development of the Internet.At present,there is a wide variety of tourist attraction services data,and tourists need to spend a lot of energy to choose their favourite attractions,which greatly affects the user experience.In order to enable users to retrieve useful information more quickly,the more widely used is the traditional decomposition based on the rating matrix,while the use of deep learning to mine the user's review information,both can improve the effectiveness of the recommendation model.Therefore,this paper proposes to optimize the recommendation model based on matrix decomposition method,using review information and deep learning techniques to further improve the accuracy of tourist attraction recommendation service.This paper proposes a new recommendation algorithm by fusing the LDA topic model with a deep learning algorithm,which firstly uses the LDA topic model to extract the topic features of users and attractions in the reviews,then multiplies the two parts of the feature matrix,and further integrates and extracts the information through a fully-connected layer;secondly uses a multi-layer perceptron as the main structure,and builds a deep learning model to The algorithm then combines the results of the above two models in a model fusion approach,so that they can reinforce each other to better learn the user-item interaction and make the final prediction for the user,achieving the goal of improving the recommendation effect.The algorithm can effectively alleviate the problems of data sparsity and inadequate model input that often occur in the application of recommendation models,thus better helping users to make decisions.The LDA-Neu MF method proposed in this paper is applied to the same travel dataset,and the best combination of parameters is selected through experiments to achieve superior results for the model,and the results are compared with those of common recommendation models.The LDA-Neu MF method proposed in this paper is applied to the same travel dataset,and the best combination of parameters is selected to achieve superior results.