Research And Application Of User Behavior In Automotive Industry Based On Telecom Big Data

With the rapid development of the Internet,the Internet has become one of the main channels for people to access information.In the meantime,the continuous infiltration of the Internet in the field of the automotive industry has given an increasing number of people the option to search and browse the relevant information on the Internet.This provides a new source of data for related research in the automotive industry.How to effectively use the massive data to analyze the users in automobile industry is of great research significance.Based on the DPI traffic data of automotive industry in Telecom big data and the machine learning method,this paper focuses on the prediction of purchase intention of users on automobile websites and the design and implementation of user portrait system in automotive industry.The main work is as follows:1.Based on DPI traffic data of Telecom operators,combined with the crawler data of the mainstream automotive websites in China,the distributed processing platform is used to complete the preprocessing and distributed storage of car user access data.2.Aiming at the prediction of purchase intention of users,this paper constructs a prediction model of prediction of purchase intention of users and proposes a feature engineering method combining user vector representation and basic statistical features.The user vector representation uses UIB-RLW(User Interaction Behavior Representation Learning by Word2vec)method to learn user access sequence.Experiments on real DPI traffic data sets validate the effectiveness of the feature engineering approach and demonstrate the effectiveness of the predictive model.3.Designing a user portrait system based on Telecom big data for automotive industry,developing a data visualization platform,elaborating on the design and acquisition of user portrait label system,presenting and inquiring the user group portraits,and completing the automotive industry insight analysis.The method proposed in this paper to process sequence data and extract user vector features by UIB-RLW(User Interaction Behavior Representation Learning by Word2vec)method is of great significance to the study of user behavior analysis in sequence data.It is aimed at the problem of predicting retained capital behavior of auto users and the constructed user retention behavior.The prediction model provides new ideas for effectively utilizing user behavior data and analyzing user behavior.The design of the user portrait system and visualization platform for the automotive industry based on Telecom big data provides great convenience for analyzing user preferences,understanding user needs,and gaining insight into user habits.