Application Of Multivariate Decision Tree Based On Rough Set In Electric Marketing Analysis

With the rapid development of domestic electric power industry and the deep-going of the market-oriented system reform process of it, the duty of the power companies has been changing from the customer management to the marketing and service. Therefore, electric marketing analysis, which decides whether the power companies can survive under the condition of the drastic market competition and get development, becomes the groundwork of the power companies. However, most traditional methods for electric marketing analysis are focus on data management or just do some simple data statistic analysis. It is very difficult to gain significant information for marketing decisions just by these methods, when facing more and more enormous database of electric marketing.Aiming at upper questions, in this paper we apply the multivariate decision tree based on rough set to electric marketing analysis. Firstly, segment the electric market according to the difference of industry and month, etc. Then apply the decision tree which is based on rough set to study the relationship between the marketing behavior and the influencing factors, such as price, precipitation, temperature etc, and quantify this relationship with certain measurement indexes. According to the results, the marketing analyst can get more veracious judgement and understanding on the characteristics of power market, and make corresponding marketing strategies.Based on the ideas, a lot of practical work has been done. Firstly, data warehouse with the theme of electric marketing analysis is built. Then the clustering method of an improved K-Means is applied to generalize the original data. Secondly, using rough set theory to reduce the number of attributes and keep only the essential attributes, and a multivariate decision tree is constructed. Finally, on the basis of the decision tree, the relationship between electric marketing and the influencing factors can be easily obtained, meanwhile we testify the results to eliminate the possible misguidance in them and guarantee the veracity of the rules.According to the practical example, it shows that the method given in this paper can acquire the rules in electric power marketing fully and quickly, and the rules may not be found easily with traditional methods. The multivariate decision tree technique based on rough set is proved helpful in aided decision-making for electric marketing.