The Prediction Of User Purchase Intention

Improvement of Internet technology, leed to the development of every field in people’s daily life, result in the prosperity of electronic commerce. In the development of electronic commerce, brands have gradually become the link between people and goods. The degree of dependence and trust people on the brand largely affect the user’s purchase intention. While the historical behavior of users is the best show of their preference on brands.The research focus on the prediction of users’ purchase intention on brand goods to achieve the goal of recommending. In this research, we have the records of users’ historical behavior on brands. By means of feature extraction and modeling, we predict that the users’ purchase intention, the brands, which tend to be purchased in a period of future.Below is the contents and innovation points of the paper:1. Data analysis and feature extraction. First, describe the data source and dataschema.Through the data analysis process, abandon the improper model and em-phasize the key point in feature extracting.Then,show all the features we extrated.2. Base model constructing and selecting.We conduct a series of experiments andconstruct the base model based on the experiment results.Also, we choose dif-ferent learning algorithms on the base model,prepare for future work.3. Model Ensemble. Based on the base model and algorithms above, we choose theproper base models,and ensemble them to improve the effect of the final model.