Consumer Behavior Analysis And Prediction Model Of E-commerce Platform

With the rapid development of information technology,we have entered the era of big data.As the young people's online shopping has become a kind of fashion,the e-commerce platform has gathered a large number of consumers'purchase data.Many e-commerce platforms make use of big data and cloud computing technologies such as Hadoop and Spark^[1]to extract useful information from high-dimensional and massive data,analyze and model online consumer behavior,and forecast consumer demand.Its main purpose lies in three aspects.On the one hand,it is used for product display,personalized recommendation and accurate delivery of advertisements;on the other hand,it is used to influence the purchase decision of users,such as the price of a product,the number of sales,the number of evaluations;finally,it is used to support the data-based decision-making of countries,regions and companies.What's more,based on the analysis results of purchasing data,consumers'consumption behaviors could be understood,the industrial structure also could be adjusted in a timely manner to make the economy coordinated,stable,and sustainable,which could benefit the society^[2].The essay aims at the problem of the accuracy of user's consumption behavior forecasting faced by the current e-commerce platform,thereby,the data mining technology is applied to the analysis and prediction of user's consumption behavior.This paper filters and combines existing data mining algorithms,and builds an algorithm suitable for analyzing and predicting user's consumption behavior by using network data.In order to ensure that the paper has practical application value,the data in this paper is the consumption data collected from Taobao.However,the data obtained from the Internet may be missed,inconsistent and dimensional inconsistencies,this paper preprocesses the original data firstly.In order to avoid the data dimension being too high,principal component analysis method is used to select mutually independent factors to achieve dimensionality reduction.In this paper,the processed data is divided into test data set and training data set.Naive Bayesian model,decision-making tree?include CART tree and conditional inference tree?model and gradient boosting decision tree model are used to fit the test set data.And then,the model that is fitted well is used to predict the data of the training set.Finally,use the Confusion Matrix to compare the accuracy of prediction results and real results of training sets of different models to provide scientific basis for e-commerce platform or enterprise decision-making,incubate emerging industries based on big data analysis,solve the major needs of users and become a new engine for promoting social progress.