Research On Industrial Enterprises Economic Prosperity Index Intelligent Prediction Model Based On Rough Set And Support Vector Machine

With the rapid development of economy, the economic situation is complicated, economic present a more uncertainty and complexity; economic forecasts received more and more attention, which to government decision-making, business and personal investment has important effect. The enterprise economic boom as "the barometer" and "alarm" of economic development, its information has the high advanced, objectivity, reliability and continuity, is monitoring the macro economy development trend and the production and operation of enterprises in one of the important means. However, the current enterprise economic variable exists between the very complex multidimensional nonlinear mapping relation, the traditional economic boom relying on the knowledge and experience by economic experts, the method need a lot of hypothesis test and model modification, and the method and the lack of nonlinear data processing and the dynamic changing environment adaptability. So economic boom analysis forecasting method turn to intelligent information processing technology, the economic boom intelligent forecast method is of great significance to the study and necessity.This paper, from the Angle of the intelligent information processing, research on the method of data processing complex systems—Rough set, and the latest, nonlinear support vector algorithm prediction method—least square support vector machine, In order to eliminate the index redundancy in economic climate index system and the nonlinear characteristics of forecasting. A new hybrid forecasting model is proposed which combining the rough sets and Least Squares Support Vector Machine, Meanwhile, for the parameters selection of LS-SVM is a NP problem, introduces an improved Particle Swarm algorithm to choose excellent parameters, increasing the precision of predict model. Finally, the prediction model is applied to the industrial enterprise prosperity index, solved the difficult problem of traditional model with complicated nonlinear data and nonlinear prediction problem.The paper is divided into four main parts:In the first section, we summarized the research situation in the world through three aspects, including economic prosperity analysis, forecasting methods and the application of intelligent methods in prosperity forecasting. In the second section, this thesis studied on the basic theory of prosperity prediction, mainly from the economic cycle, the prosperity index establishment, and the prediction aspects to describe the basic theory of prosperity index and the classics method of prosperity prediction, which laid a solid foundation for forecasting.In the third section, the procedure is as follows:First, introduces the basic theory of rough set method and attribute reduction algorithm, and its applying to economic prosperity index system construction and the index screening, solving complex data system information redundancy; second, introduces the basic theory of SVM and the research status. This paper focuses on the LS-SVM algorithm model theory, in view of the LS-SVM parameters are difficult to be determined, based on the improved particle swarm optimization algorithm is proposed so that improve the prediction effect. At last, with the advantage analyze of RS and LS-SVM, a novel PSO-LSSVM predict model Based on RS was constructed, and the model theory, method and procedure are analyzed.The fourth section, the construction of prediction model applied to industrial business prosperity index prediction for empirical simulation analysis, and compared the prediction effect with standard LS-SVM and BP neural network; the results of empirical simulation verify the effectiveness of the proposed method.In the last part of this thesis, that is, sum up and outlook section, gave a brief summary of the work and future work direction of this thesis.