| Case-Based Reasoning (CBR) method is a theory with promising application prospect and is widely used in many fields. It's a summarization of past experience as well as a reference of past works. It can overcome some defaults in current knowledge processing system and promote system's capability of reasoning. CBR for Cotton Textile Crafts and Arts Expert System (CTES) is a practical application of the theory. The research is to develop new crafts and arts in cotton textile field. Our country is a big country in cotton textile field, but the technology is less than satisfactory, the construction of this system for solve this problem has positive function to a lot of medium and small-scale cotton textile enterprises.Case-Base Reasoning is like the way of human's thinking and dealing with questioning. The main work is found the most fitness method from comparing the target case and the case in case library. Equally, adaptation was down when the old case is not completely in line. It has much more respects when it is used in specific fields. This paper mainly studies CBR for cotton textile crafts and arts expert system, some works was done as follows:At first the paper introduce the background and the significance of the study and current situation of studying both at home and abroad in Artificial Intelligence (AI), Expert System(ES), Case-Based Reasoning(CBR) and cotton textile crafts and arts.Secondly, the paper has a roughly introduction on CBR. What is the meaning of CBR? What is the course and model of CBR? And what is the characteristic of CBR compare with other reasoning, such as Information Retrieval (IR), Rule-Based Reasoning (RBR), and so on.Thirdly, the paper analyzes procedure and characteristic of the cotton textile crafts and arts, and has a total design to the system. Such as the division of module, design of database and code, input and output designing etc.Fourthly, the key technology for realize the system was carry on discussion, including knowledge representation, case retrieve, case match, case revise and case reuse, etc. Retrieve of case is mainly discussed. There are many methods for case retrieve in CBR expert system. One of these methods that are used widely is near neighbor method. It gives a weight to the attributes of cases, and calculates the distance of two cases for compare which is cases nearest neighbor. But this kind of method is time consuming in a situation when the case library is huger more, so how optimize and improve this methodgo on the discussion thesis.There are two chief questions. One is the time of case retrieval, and the other is accuracy. For the first question, the combination of nearest neighbor method and the screening method in different levels combine with rule according to some tactic is employed to develop the retrieval efficiency and reduce the process of retrieval. For the second question, the genetic algorithms is used for improve accuracy of retrieval by alter the weight.At last, the paper introduces the realization of the system and verifies the execution of a prototype system. And summarizes all of the work, expects to future.It is testified by experiment that CTES prototype system can reason effectively according to the input of the users and educe appropriate crafts and arts. |
PDF Full Text Request