Multi-attribute Car Assessment System Based On Affective Ontology Tree

As an important expenditure, buying cars is becoming a more and more difficult and important choice so that people always make a decision carefully. A problem people faced is that how to evaluate the multiple attributes reviews information based on language commentary.Ontology knowledge database is an available resource.This article solves the problem with mathematical methods and information theory. At first, we get the matrix by matching the affective ontology tree. Then compute the weight of attributes using entropy weight method. The default of an attribute is common in fact. Dempster-Shafer evidence theory is an effective method which could compute the default of attributes by using the Dempster-Shafer evidence method. At last, we get the score of the information given by people. What we do in this article is something as the following information.(1) The conversion of affective ontology treeAccording to jena and pellet, we construct a tree structure through which we could understand more about the ontology tree. We design a way in which we could transfer a tree from protege to eclipse.The tree will be used to match the information of the reviews.(2) Compute the weight of attributesWe get the matrix called review matrix by matching the affective ontology tree. In this matrix, attributes of cars are different from one to another which means that the value of the weight is distinct.In this article we conduct a method which called entropy weight method to compute the weight of some attribute.Then we receive the weighted matrix.(3) Decision method based on D-S evidence theoryUsing the associative and commutative law, we conduct the matrix by evidence combination rules through which we combine two evidences. Then we combine another new two evidences. At last we receive the result of the comments. This theory can perform well when there is default attributes in the given comments.(4) Multi-attribute car assessment systemMulti-attribute car assessment system based on affective ontology tree contains four crucial parts, displaying the ontology tree, matching the tree, computing entropy weight, using dempster-shafer evidence theory to mass the result. Input comments about cars, this system will transfer these comments from words to figures, then compute weight and mass the result.