Studies and Applications of Inegrity Based on Flexible Logic
|School||Beijing University of Posts and Telecommunications|
|Course||Signal and Information Processing|
|Keywords||Integrity Logic Universal Conjunctive Operational Model Universal Disjunctive Operational Model Shape Features Feature Extraction|
One of the core issues in intelligent information processing is how to deal with several kind of uncertainty reasoning, but the uncertainty is external behavior of objective things, which exist and changing are confined by the inner dialectical contradiction of objective thing, and the important goal for logic development is building flexible logic (which is the mathematic dialectical logic in logical field). It has long been found much non-standard logic lost some important properties of standard logic in practical application, its logical reasoning lost "credibility" although it is still "reliability" and "completeness". In order to avoid distortion of information during processing, professor he gives concept of sound logic According to transmission and use of Information in the practical application in 2008. According to transmission and use of information in the practical process, professor He hua-can gives the concepts of sound logic in order to ensure accurate of information in the process. Integrity about logic is studied based on the analysis of development situation of continuous-valued logic. And main achievements are as following:1) Integrity about 0-level universal operation model is proved.On the base of analysis reliability and completeness of binary logic and fuzzy logic, only reliability and completeness are not enough in continuous valued logic to ensure that information is distorted in reasoning process. Therefore, this paper introduces integrity to further ensure credibility in reasoning process of continuous valued logic reasoning. This paper proves that logical system PC(T) (h∈(0,0.5]) based on 0-level universal operation models is an integrity logical system. An integrity propositional logic PC(T) (h∈(0.5,0.75]) based on 0-level universal operation models is built. Moreover, we prove that propositional logic system PC(T) (h∈(0.75,1]) based on 0-level universal operation models is an integrity logical system for P=0,1. An integrity propositional logic PC(T) (h∈(0.75,1]) based on 0-level universal operation models is given for P≠0,1. In the practical application of fuzzy decision tree, algorithm of matching operator in the fuzzy decision tree is improved based on integrity of flexible logic and provides the principle to select the matching operator.2) The relationship of mutual information I(X;Y) and generalized correlative coefficient h is studied, A fuzzy reasoning model based on universal combination operation model is proposed.This paper discusses relationship of mutual information I(X;Y) and general correlation coefficient h and gives the the corresponding rules of them, the optimal matching operator is chosed to complete fuzzy decision according to mutual information between candidate attributes. The relationship of mutual information I(X;Y) between attributes and generalized correlative coefficient h provides the principle to select the matching operator. According to the results of experiment, it is more reasonable to enhance classified precision effectively. This paper’s method has a certain application value.3) Brain CT image classification by fuzzy decision tree based on universal combination operation model.Segmentation in brain CT image and feature extraction are studied combining shape feature. First, region growth algorithm is improved according to characteristics of grayscale distribution. According to the results of experiment, it shows that the can segment interesting semantic region clearly. It is helpful for subsequent feature extraction. Second, according to position relation with different parts of brain CT and morph feature of the pathological region, a method is is proposed based on feature extraction of tree structure considering the adjacent parts of tree structure guided medical prior knowledge. The experimental results show that average classification accuracy improves 27.3% and 11.9% compared to the whole picture and single part.