The Implementation Of Data Mining Technology In Single-Source Survey Data

With the development of market economy, the market research company plays an important role in helping the business enterprise make the management decision and establish the marketing plan. They help enterprise understand the market status, obtain the market information and improve the effects of the marketing.Born in UK in 1966, the single-source survey is one of important formats of the market research company's serves. Not like tradition format which provides the service for each customer, the single-source survey concentrates manpower and material resources to establish a periodical renewal database including demography, product consume, media touch and life style information of people. Then this database is provided to many customers (including advertising agency and enterprise) to support management decision-making and marketing decision-making. For the concentration of manpower and material resources, the database of the single-source survey has the character of wide information and huge data. So the analysis of the database needs the special software.Moreover, the major functions used in the single-source survey database analysis software are cross table, P-Map, correspondence and media planer etc. Using these methods to describe the character of target customer has the problem of huge workload, bad information finding ability and personnel ability influence. However, the data mining technology can solve these problems. This paper uses data mining technology to design and establish the customer character-describing model.The major contributions of this thesis lie in:1. After analysis of the characters of data mining technology and thesingle-source survey database, we select appropriate knowledge discovery technology to support the single-source survey research.2. Designing the decision tree based model, we accelerated the analysis speed and knowledge discovery ability of the customer's character analysis.3. Using the open-source system Weka, we implement the customer's character analysis.