Study On The Application Of The Fuzzy Decision Tree And System Design Implement

The decision-making, selected by the decision-makers who utilize their domainknowledge and dominate some controllable variables in order to achieve specific purpose, is aprocess of maximizing the utility of the program. Nowadays, with the massive informationand the greater uncertainty presented by knowledge, the decision-making is facing a morecomplex surroundings. Therefore, the study of the intelligent decision-making system is amatter of great significance.In the field of data mining and knowledge discovery, the range of some variable showssome kind of the fuzziness. Thus, the establishment of fussy decision tree is meaningful in theanalysis. The fussy decision tree is a combined technology, which is composed by thedecision tree and the approximate reasoning offered by the appearance of the fuzziness,aiming at taking advantages of the good points of the two components, namely theuniversality of the application of the fuzzy decision tree and the understandability of theknowledge, as well as the ability of processing the imprecise and uncertain information of thefussy representation. The combination of the fuzzy set theory and the decision tree solvessome problems which cannot be worked out by the traditional decision tree.The present thesis mainly studies on the simulative implement and application of thefuzzy decision tree, and the details are as follows:1. A comparative analysis between fuzzy decision tree theory and traditional tree iscompleted after studying on the decision theory and related algorithm, including thecomparative analysis of the merits and demerits of the process of establishing tree, theselection of the attribute of the category, the dominance of the growth mechanism and therule extraction. The fussy decision tree can deal with the discrete attribute, continuousattribute and fussy attribute. Fuzzy set and approximate reasoning can tackle the noisy,inconsistent and incomplete data which is more precise than the tradition decision tree.2. With the development of the loan service of the domestic commercial bank, there isan increasing trend that the expansion of all kinds of loan size is developing year by year.Particularly, the personal credit business is becoming a more and more important operatingway of increasing income of the major financial institutions. Gradually, the development ofpersonal credit business will be the strategic business of the commercial bank. At present, thedomestic personal credit business adopts the consumer scoring as the credit evaluation beforethe loan. This paper introduces the fussy decision tree into the consumer classification of personal loan credit assessment. Through applying the assistant system of fussy decision tree,it testifies and makes a fussy classification over the data which are from the practical creditloan data, and gets a more precise practical result than that of traditional decision tree.3. This study constructs the fussy decision tree system which develops as the learningmechanism and centers on the rules. The main function of the fussy decision includes thefussy data processing, the generation of the fussy decision tree, the rule extraction of the fussydecision tree and the data check of the fussy decision tree. The system, based on theMin-Ambiguity algorithm, is featured as the understandability and swift construction and soon.