Electric Marketing Analysis Decision Support System Research And Design

With the rapid development and further deregulation of electric power industry, the responsibilities of electric power companies have been changing from user management to marketing and service, as a result, electric marketing analysis becomes the primary operation of power companies. However, so far, the actual marketing systems just manage the data information, the quantity of which is various, so it becomes a hot topic how to find the regulations to support the decision with computer technologies. This paper integrates the marketing analysis with Decision Support System (DSS) to build the Electric Marketing Analysis Decision Support System (EMADSS) that adapts well to the electric industry. Because model-base is an important criterion to evaluate DSS, different models, which are based on Data Mining (DM) and Multi-Statistics theories, are applied in this paper. Firstly, Association Rule (AR) and Decision Tree (DT) in DM are presented to analyze the trait and potential of the electric selling market. In order to eliminate disadvantages of buzz handling in DT, Rough Set (RS) is integrated with DT to build an algorithm that can explore the market potential effectively. Secondly, many models are established for load forecasting and the sensitivity analysis with Multi-statistics theories. Because of the complicacy of the integrated sensitivity analysis, the method that combines Principal Component Analysis (PCA) with non-linear function transformation is proposed. PCA is introduced to erase the coupling relations among the factors in the electric market. Then the function transformation is applied to gain the complicated functions of electric consumable distribution. Lastly, based on the above theories, EMADSS can provide valuable suggestions to exploit electric market, which can make the system more practical. According to the above models, EMADSS software is researched and developed. It is designed to be hybrid architecture of C/S and B/S. The algorithms and database of system employ COM and data warehouse technology, respectively. The design style assures stability and reliability of the system.