Some Combination Forecasting Models Of IOWA Operator Based On Uncertain Information

The combination forecasting is one of the important research methods in the field of prediction and has been widely used in the fields of economy,society and management.It can not only improve the accuracy of prediction,but also can disperse the prediction risk caused by the lack of information contained in each single prediction method.The key of combination forecasting method is to determine the weighting coefficient of each single prediction method.In this process,we should consider not only the information itself obtained by single forecasting methods,but also the mutual influence between the data.In addition,with the increase of the complexity and uncertainty of social system,the data which describes the system often has some ambiguity,so we need to introduce the concept of fuzzy number to forecast the complex system.This dissertation discusses some combination prediction of fuzzy systems under uncertain information.And this dissertation is divided into the following six chapters.The first chapter mainly introduces the research background and significance of the combination forecasting method.It also introduces some main research results in this dissertation.In chapter two,the concept of interval number and triangular fuzzy number and some algorithms are introduced,and the definition of IOWA operator is introduced.The third chapter proposes a new conception of interval prediction accuracy,which uses interval number to express the prediction accuracy,and rank them with possibility degree.The indexes of prediction accuracy are utilized as induced variables,a variable weight combination forecasting model is established based on interval prediction accuracy and induced ordered weighting averaging(IOWA)operator.An example is illustrated to show that the model is valid.In the end,the sensitivity analysis of the parameters is done in the model.In chapter four,when we use combination forecasting method to forecast triangular fuzzy number,we should consider the average prediction accuracy between the prediction number and the actual number as the induced variables,then,we introduce the concept of IOWA operator to reorder and weight the triangular fuzzy numbers which are forecasted by each single prediction method.Next,we use vector angle cosine to solve the optimal weights,so we establish the fuzzy combination forecasting model based on IOWA operator and vector angle cosine.An example is illustrated to show that the model in valid.In the end,the sensitivity analysis of the parameters is done in the model.Based on the fourth chapter,the fifth chapter considers the traditional combination forecast for triangular fuzzy numbers is to combine their three endpoints.But this method will cause the disorder of three endpoints for the triangular fuzzy numbers and affect the prediction information integrity.This paper introduces the concepts of the area covered by the corresponding membership function and gravity center of triangular fuzzy number.The indexes of prediction accuracy of area and gravity center for the triangular fuzzy number sequence are redefined.The IOWA operator is utilized to build the multi-objective combination forecasting model based on the correlation coefficients of corresponding area sequence and gravity center sequence of the triangle fuzzy numbers,respectively,in which the importance parameter is introduced to transform it into single objective programming model.An example is illustrated to show that the model is valid.In the end,the sensitivity analysis of the importance parameters is done in the model.In chapter six,the main research results of this paper are summarized,and the future research of fuzzy combination forecasting is prospected.