Recommendation Algorithm and Application via Big Data Mining

Nowadays, the amount of data is increasing second by second, which implies huge commercial value in the developing economy. How to mine effective information from the data ocean has become a hot topic in various fields. Here we chose the ml-latest-small data from MovieLens data set to realize the efficient recommendation to the consumers. The K-means clustering algorithm in machine learning technique and the user-based collaborative filtering algorithm are reconstructed, and the cluster-based collaborative filtering recommendation system is implemented. Empirical analysis shows that, the model training time is shortened from one minute to 10s after switching the data storage method and clustering to 5 categories, which greatly improves the efficiency of the entire recommendation system.