Study On Structural Theory And Algorithm Of Power Sales Decision Support Systems Based On Data Mining

With the stepwise development and improvement of China electric power market, the power industry will be changed from "seller's market" to "buyer's market", which will give our power industry a great influence. As independent market main bodies, the power enterprises's operating goal will changed to pursuing the maximum benefits of the enterprise, and their working emphasis will be converted from power's generating, transporting and distributing to the market development, power demand's management and so on.With the mechanism's conversion, the traditional MIS (for example, power use MIS) has not already satisfied power enterprise's needs. How to build the imformation system to satisfy the needs of our power sale's decision has become our urgent affairs.The data mining(DM) technique which is used to discover the potential relationships and rules in vast dataware is the product of combining the artifitial intelligence and dataware techinique, and has already become an important method for intelligent decision and way for getting decision knowledge, and has the important applied research values in decision support systems.The new structural frame designing theory of power sale decision support system is presented based on careful demand analysis and design of power sale decision support system. The new frame has the features of decision problem's leading function and the combination of dataware(DW) techniques, on-line analytical processing (OLAP) techniques and data mining techiniques, which can preferably satisfy the needs of power sale decision support.The OLAP technique based on DW is one of the important techniques in power sales aided decision support. A datahouse scheme on powe sales is presented based on careful analysis of datahouse; The OLAP analytical contents and methods are studied; The muti dimensional analysis of power quantity and rate is designed and carried out by use of BusinessObject OLAP tool, including universe design, general query report form design, slicing, drlling, circumrotating and so on.The DM and OLAP are the critical techniques of data analysis, the multi dimensional data mining model combining both techniques can enhance the performance and effects of data analysis. As far as the choosing method of mining space in multi dimensinal mining model, a new ANN strcture and algorithm are posed, which avoids the problemof complicated nonlinear modeling and has less calculating than general ANN variable method.In view of the important effect of clustering analysis in data mining, to solve the problem of determining clustering number, the clustering rules and its curve are studied carefully; A kind of self-adaptation clustering ANN is presented based on SOFM ANN, which can automatically determine the clustering number. Based on pratical sales data, the time feature analysis of power user consumption are carried out by using the self-adaptation clustering ANN, whose conclusion has the impotant referenced values for adjusting power price correspondingly and arrange power producing reasonably.The achievements of the dissertation have the important referenced value for the scheme designs and realization of powe sales decision support system of power enterprise under power market.