User Behavior Data Analysis of Taobao Online Based on Flink-Based K-Means Algorithm

On the k-value sensitive effect of K-means algorithm in data clustering, based on the characteristics of density distribution, the kernel density selection scheme is adopted to improve the algorithm. For massive data, the improved algorithm is parallelized based on Flink platform. In the practical application of mobile e-commerce, experiments are repeated in serial mode and parallel mode respectively, and the improved algorithm and K-means algorithm are used for specific comparison. The results show that the improved algorithm will also have good clustering effect in real life. It can be seen that the improved algorithm not only has accuracy and efficiency, but also has practical significance.