The Analysis And Application Of Automobile Sales Status Based On Data Mining Technology

Under the background of Internet plus and big data, auto enterprise management and marketing are more dependent on the data collection and analysis of the customer behavior, preferences, and support every link of company sailing such as the potential market development and customer support maintenance. Potential auto customer analysis is an important technology of the brand and customer relationship, which is mainly through the exhibition, store call customer visit and shop visit, WEB data mining of on-line analysis of customer behavior and preferences to mining potential customers and effective support for market promotion effect. Collected from a variety of sources of information exists the basic characteristic of the big data which is large data capacity, unstructured, strong real-time. It brings a challenge for data analysis, and leads to the accuracy problem of the mining results. For the above problem, through the analysis of the data without a mentor, the customer clustering analysis is on research. It is solved the accurate potential customers to identify. The method based on principal component analysis (PCA) and two stages of clustering analysis of the analysis of the potential auto customer is proposed. Moreover, in order to analysis and make use of the existing customers quickly and efficiently, the improved Hash fast attribute reduction algorithm based automotive customer analysis model is presented. The main work of this thesis is as follows:1. Based on principal component analysis (PCA) and two-stage clustering analysis in the potential customer development and management research are proposed. According to automotive customers choose their own characteristics, the cluster analysis can automatically data mining, which has low operating cost, occupy less memory resources and the characteristics of simple and feasible. Using principal component analysis can extract the car customer classification of the main influence factors, and the two stage cluster analysis method is made used of the analysis potential auto customer.2. A hash quick attribute reduction algorithm and its characteristics and existing defects are analyzed. An improved algorithm is put forward, which could improve the efficiency of reduction algorithm to make further reduce computational complexity. It can improve the efficiency of up to 20?30% in many experiments. Computational complexity and efficiency in t the system, the superiority of the improved algorithm is summarized.3. Intelligent analysis platform is developed. It makes use of historical sales customer database information and previous research data information to generate the decision tree to help sales provide different sales strategy according to different customers. It could enhance the customer loyalty and satisfaction strategy in the automobile sales and service. The aim of the system is to direct sales personnel to recommend different car configuration according to the customer's information through the attribute reduction results of previous sales database information which has a certain guiding value for better customer service practice in the car sales industry.