Fuzzy Statistic Theory And Its Application In Economy Forecasting And Decision Making

In this thesis, we firstly introduce the relative theory of fuzzy statistic and fuzzy decision making, and then the applications of fuzzy statistic into fuzzy decision and fuzzy classification, clustering, identification have been studied in the following Chapters.Firstly, the law of large number of fuzzy numbers sequence in R is discussed. In the second Chapter, a fuzzy Markov chain model with the use of fuzzy transition matrix is presented to forecast the trend of bank loan ,due to Markov chain's characteristic that future state of system is only relate to present state and the system state at some time may submit to divergent fuzzy state. In Chapter 3, fuzzy comprehensive evaluation is applied in fuzzy statistical analysis firstly. And then a similar approach degree method is proposed to statistical decision making by integrating fuzzy clustering with grey association analysis. Based on fuzzy evidence theory and possibility theory, the model of indefinite Bayes statistical decision making in the face of external information resource with fuzzy belief structure is discussed in Chapter 4.In the area of intelligence, vague set is being used more widely than fuzzy set. In Chapter 5, a new Bayaes statistical model based on vague set is presented for decision making. In this model, the probability of appearance, disappearance of nature situations and that that we do not know whether it will appear are processed respectively.On the basis of L.zadeh' s composite principle of fuzzy reasoning rule and the Chen' s method of character expansion of fuzzy reasoning, a generalized mathematical model of fuzzy character compound reasoning and forecasting is presented in Chapter 6..Based on fuzzy set theory, we propose three statistical methods, i.e, fuzzy select matrix, fuzzy comprehensive evaluation and fuzzy relative membership, to resolve multiple object decision making with fuzzy information in Chapter7.In the final Chapter, a similar approach degree is presented and then applied in fuzzy model decision classification, based on three kinds of measures of approach degree between two fuzzy sets, i.e, Hamming distance, fuzzy nearness degree and grey association degree.