Font Size: a A A

Cognitive Pattern Recognition Theory And Research On Chinese Character Intelligent Information

Posted on:2011-04-05Degree:DoctorType:Dissertation
Country:ChinaCandidate:M Y LiuFull Text:PDF
GTID:1118360308463658Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Pattern recognition is a basic cognitive ability or intelligence of human. The existing pattern recognition theory and method of things with the human cognitive processes are quite different and have some deficiencies on learning ability. This paper combines cognitive science, psychology and informatics to study the cognitive pattern recognition based on the prototype matching theory. The theory and technology of Chinese character intelligent formation was studied and discussed in this paper based on cognitive pattern recognition.The theory of cognitive pattern recognition was studied and discussed deeply. The perception perceptual contancy theory of congnitive psychology, component identification theory and prototype matching theory were introduced into the machine pattern recognition, pattern invariable theory, pattern composition theory and cognitive matching theory were studied and discussed, and the structural framework of cognitive pattern recognition theory was proposed.Take the cognition pattern recognition theory as a foundation, Chinese character intelligent formation without library was studied and discussed deeply. Based on the traditional Chinese character creation theory, the mathematical model of Chinese character intelligence formation was set up; the framework of Chinese character intelligent formation was proposed. Sumed up 1085 prototypes of Chinese characters through the analysis and abstraction of all the components in Chinese character set GB18030-2005. The relationship between the combinations of Chinese character prototype is known as the structure of Chinese character. 18 kinds of Chinese character structure were proposed through the split and structure analysis of Chinese character set GB18030-2005. Thus enriched and improved the theory of Chinese character intelligent formation.In order to make the Chinese character intelligent formation theory into practice, this paper researched on the key technologies of Chinese character intelligent formation deeply.Encoding is the first key technologies of Chinese character intelligent formation. The coding method was researched through some experiments. On the basis of basic elements plus structures, enriched the encoding method based on global priority local of topological perception theory. Encoding experimental platform was developed and the encoding of Chinese character set GB18030-2005 which contains a total of more than seventy thousand Chinese characters was completed. Experiments show that the method consistentes with the recognition laws that local after global of characters, integration with the primary basis for language education, and re-encoding rate is zero.The mapping method from basic elements to the specific components of Chinese character was studied, a Chinese character can be formed by geometric transformtion of basic elements in according to corresponding knowledge. From the characteristics of basic elements mapping, the geometric transform model of basic elements was built. Put emphasis on how to acquire the mapping knowledge of basic elements, two methods that based on region and feature extraction algorithm for acquiring the affine transform coefficients were proposed. Has carried on the mapping knowledge acquisition and form characters experiments that choose Chinese character set GB18030-2005 as experiment object.The organization and representation of Chinese character intelligent formation was studied, the knowledge library of Chinese character intelligent formation are including the structure knowledge, basic elements knowledge, coding knowledge, and basic elements maping knowledge. This paper proposed represents the above knowledge with the semantic network. The semantic networks were built which take the knowledge of Chinese character set GB2312-80, GB18030-2000 and GB18030-2005 as a knowledge source respectively, the storage and reasoning methods of the semantic network were proposed, and the performance evaluation method of the semantic network was proposed and has carried on the analysis. On this basis, according to the needs of Chinese character intelligence formation, the experimental platform of Chinese character intelligent formation was designed and developed.
Keywords/Search Tags:cognitive pattern recognition, Chinese character intelligent formation, basic element, Chinese character structure, knowledge acquisition, knowledge representation
PDF Full Text Request
Related items