Log Data Mining of User Purchase Behavior Based on Distributed Intelligent Optimization Algorithm

As the development of e-commerce becomes more and more intelligent, higher requirements have been put forward for the algorithms controlling e-commerce operations. However, the current e-commerce operation is not timely and accurate enough to update the purchase data and statistics, resulting in cost consumption and revenue is not proportional, and can not accurately meet the user favorite. To speed up the collection of user purchase behavior data and improve the revenue of e-commerce operations, the study introduces adaptive degree values based on a distributed computing framework combined with a topological structure. The computing framework is used to speed up the calculation and convergence of user data, and the topology is responsible for classifying the data in the dataset and calculating the optimal location. In the classification accuracy experiment, the accuracy of the improved algorithm was above 94% and up to 98%. In the stability experiment, compared with other algorithms, the stability of the improved algorithm was improved by 81.2%. In the simulation experiment, the overlap between the noise value of beauty search 2000-2700 and the noise value of clothing matching 2000-2500 in the shopping platform was large. Therefore, there was a correlation between the user's search for clothing collocation and the beauty search. In summary, the performance of the improved algorithm is superior in terms of stability, accuracy and application error. Therefore, the study of the improved algorithm has a better application for data mining of user purchase behavior.