Research On Online Consumer Behavior Based On Machine Learning Theory

With the rapid development of online shopping platform,on the one hand,the growing maturity of the Internet has facilitated people's life and improved people's quality of life,on the other hand,the resulting huge data implies huge value to be developed.The research of online data not only has great social and commercial value,but also has great academic research value.The behavior of online consumers is complex,and there are many analysis angles.How to mine effective data from massive data,and what kind of analysis angle is more efficient and meaningful,which are always the problems that researchers are thinking about and improving.With the emergence of more data,the tools for data analysis have been developed.Since the 1950 s,the use of machine learning has been fully released and applied in the era of big data.Machine learning in different periods has completed different missions,from machine learning to mathematical theory support,from the beginning of statistical machine learning to the improvement of deep learning,machine learning has a lot to dig and use.Artificial neural network imitates the principle of biological thinking,improves itself in the process of continuous learning,and greatly strengthens the ability of data processing and analysis.So how to understand the application of machine learning in the new era? How to use technology to find answers to confusing everyday questions in data? Based on the real data of shopping platform,this paper establishes problems from differe nt angles and provides corresponding solutions.It uses BP feed-forward neural network to process historical data to predict whether online consumers' purchase behavior occurs or not,and uses LSTM cyclic neural network to analyze index accumulation to predict the amount of purchase.Starting from two practical examples,this paper analyzes and deals with the BP Back Propagation neural network and LSTM cyclic neural network,so as to further understand the role and charm of relevant theoretical knowledge.The prediction of online consumers' purchase behavior is to grasp the next behavior trend of consumers through the existing behavior data,and judge the final behavior of online consumers through these data.On the one hand,the analysis and prediction of online consumers' purchase behavior should understand the characteristics of commodities,purchase quantity and corresponding consumer groups;on the other hand,the purchase habits of each consumer should be considered.