New Intelligent Pattern Classifiers based on Human Deep-thinking

The development of new intelligent classifiers based on human thinking logic is a challenging research topic in the area of pattern classification. Currently, most classifiers are designed by using mathematical optimization-based techniques, rather than human thinking and reasoning. This thesis focuses on the development of new intelligent classifiers, based on human deep-thinking logic, to approximate the optimal design process in conventional classifier designs and to avoid the complex and time-consuming mathematical optimization-based trainings. This research will greatly enhance the learning efficiency and improve classification performance.