Product Recommendation Algorithm Based On Data Analysis In Mobile E-commerce

With the continuous development of the Internet and the transformation of people's consumption concept,the e-commerce platform has become people's primary shopping way. In that case,personalized recommendation system becomes the main means to help users quickly find favorite goods and to help platform sell goods as well.The collaborative filtering recommendation algorithm is the most widely used in the recommendation system, the main idea of this algorithm is to find the similar neighbor set of the target user and recommend the preference of the neighbor users as the recommendation item to the target user.However, with the rapid development of the mobile Internet,traditional collaborative filtering recommendation algorithms are facing more challenges. First of all, in the mobile Internet environment,e-commerce platform can get more information about the user, such as user location context information. Those information has great impact on the similarity calculation between users in recommendation algorithm.Secondly, the mobile Internet has brought more data to the recommendation algorithm, the real-time performance of the recommendation algorithm can not be effectively guaranteed in the big data environment. In this paper, we carried on thorough research and analysis and proposed a corresponding solution to the two problems that the recommendation algorithms are faced with in the mobile Internet environment.For the problem of user location context information affects the similarity calculation in the recommendation algorithm in the mobile Internet environment, this paper puts forward a Collaborative FilteringAlgorithm Based on Location Context (CFLC). This algorithm takes the user location information as the key factor and combines with the improved cosine similarity calculation method, proposed the similarity calculation method based on the user context location to calculates the similarity between the users, and then searches for nearest neighbor set of the target user, finally recommend the interests and preference of the neighbor users to the target user.For the severe overload problem of traditional collaborative filtering recommendation algorithm caused by huge amount of data from e-commerce platform in the mobile Internet environment, this paper proposes a Collaborative Filtering Algorithm Based on Parallelized k-medoids Clustering (CFPKM). This algorithm uses the improved parallel iteration k-medoids clustering algorithm to preprocess the user-item scoring matrix in the recommendation algorithm to classify the users into k clusters. Then, it calculates the similarity between the target user and the other users in the same cluster, improves the accuracy and recommendation speed of the whole recommendation algorithm by reducing data space.Finally, by simulation of the proposed algorithm and comparison with the traditional collaborative filtering recommendation algorithm, it is proved that the recommendation algorithm is more effective and accurate when considering the user location context information and using the clustering method to preprocess data.