Research On Classification Prediction Model Based On Online Purchase Behavior

In recent years,the rapid development of Internet technology has promoted the growth of e-commerce,and the increase in the use of e-commerce has created a huge market potential.With the popularity of the Internet and online shopping,the appearance of shopping online has enabled consumers to enjoy the convenience of staying at home.The varieties of products,cheap prices,and convenient price comparison have made it the first choice for most consumers.In order to meet the demand of consumers,major e-commerce platforms need to dig out consumers’ shopping habits and consumption preferences in the growing shopping log.This paper learns the implicit purchase patterns from a large number of consumers’ historical network purchasing behavior data to obtain a model.When new consumers’ purchase behavior data is input into the model,it can predict consumers’ purchasing behavior.This paper first theoretically explains the principles of normal traditional machine learning algorithms and deep learning algorithms,and briefly describes the indicators and methods of evaluation models.From the classification of consumer purchase behavior of the comparison of the forecasting model and deep neural network prediction model classification research,we carry out the theoretical analysis,model building and algorithm implementation of these two aspects,then the classified prediction model is further optimized and improved,from the algorithm and the prediction results this paper summarizes the two aspects of each model,draw the conclusion.In the machine learning models,some multi-fusion models are proposed for the shortcomings of the traditional classification prediction model.The traditional classification prediction model is fused by using the Stacking method and Boosting method to obtain a new classification prediction model.The multi-fusion models are used to verify for the data in this paper has a better prediction effect.In the deep learning model,an improved model is proposed based on the deep neural network model,which is optimized from the perspective of data and algorithms.The positive and negative sample ratio and the activation function of the deep neural network are adjusted respectively.The result proves that the optimized deep neural network model has a better prediction result than machine learning models and traditional deep neural network model.The research in this paper is data-driven and uses different models to demonstrate the feasibility of using previous consumers’ behavior data and predicting future consumers’purchase behavior.The classification prediction model proposed in this paper can be used in the recommendation system of e-commerce platforms to improve the conversion rate.