Research Of Intelligent Infancial Decision Support System Based On The Data Mining

With the deep development of instruction of information in company, the company stimulatesan increasing number of business data. At the same time, continually changing market condition andserious competition will strengthen the willing of managers to pursue further information.Intellectual financial strategy supporting system is the concrete application in financial field whichis formed after the integration of traditional strategy supporting system and artificial intellectualexpert system. The application of intellectual financial strategy supporting system can helpcompany deal with data, such as classifying, arranging, processing and analyzing. This process willtransform the history data, which can’t work further, to information in knowledge database, whichcan be used by the senior managers. However, it is not concrete and detailed enough when it istalked about the application of data mining technology in the majority of research on the intellectualfinancial strategy supporting system. It is just show the model database and the instruction ofknowledge database in general term. Also, it does not take the real condition of the company intoaccount and the maneuverability is not strong.To deal with the problem happened in the real application, we do fieldwork in ZT group,integrate the condition of instructions of group information and the need of managers, analyze thegoal of system and the function which is needed to be fulfilled, make a concrete definition ofdifferent module’s function and use system instruction, which concludes four levels, to complete thegeneral designation of ZT group intellectual financial strategy supporting system and form fourthemes, including financial forecasting and decision, financial plan and control, financial analysisand report, evaluation. Because several function should be fulfilled by advanced data miningtechnology, this paper do separate research on three main data mining technology in this system. 1)Clustering. After comparing blurred clustering and grey clustering, we decide to use the latter andmake it into the application of stock forecasting and decision. 2) Decision tree technology.According to the characteristic of ZT group’s program investment and the functional need, I usedecision tree method to solve the problem happened in investment program. 3) Conjunction rulemining technology. After analyzing the disadvantage of classical conjunction rule mining algorithm,we use function mapping to transform the scan of database to scan of matrix and transform thecalculation of repetitive rate to the calculation of function inner product. So that it can improve theefficiency of calculation. Finally, we use JAVA to implement the algorithm and check the functionof the improvement by compare the former calculation efficiency and the latter one. Also, according to the need of ZT group, we apply the improved algorithm to the mining which investment decisionsubsystem does to stocks rule.At the last part of this paper, we take the investment decision subsystem as an instance andanalyze the instruction and exploitation of ZT group intellectual financial decision supportingsystem. At first, we define the goal, functional structure of the subsystem and the data miningtechnology which it applies, including decision tree technology and conjunction rule miningtechnology. Furthermore, we do general structural analysis to the subsystem and make the wholeprocess to four levels, which are basal data level, data store level, application logic level and resultemergence level. Then we define the module of the system and the function needed to realize inrespective module. At last, we apply to JAVA to fulfill the exploitation of the investment decisionsubsystem.