Forecast Analysis of Target User Based on Data Mining

Abstract

In the increasingly fierce market competition under the data-driven environment, companies began to focus on precision marketing to reduce costs and increase marketing efficiency and market competitiveness. In the past, most studies focused on the absolute accuracy of customer purchase intention, while little attention was paid to the accuracy of prediction methods. Based on the analysis of the advantages and disadvantages of data mining algorithms, this paper uses different algorithms to compare users' characteristics, preferences and other information implied in the purchase information. It also forecasts and analyzes the actual purchase situation of potential target customers. The results show that the prediction results of the decision tree algorithm are better than the clustering analysis and Naive Bayes algorithm, and the degree of improvement is even greater. At the age of 45-45 and commuting distance is 1-2 kilometers, there is a greater possibility of buying a replacement scooter without a car or a car group, thus providing personalized recommendation for customers to improve the quality of marketing.