Case Clustering Study Of The Externalization Of Tacit Knowledge Based On Improved FCM Algorithm

With the continuous progress of science and technology and the rapid development of the Internet,companies continue to grow and develop.The accumulation of knowledge is also increasing.How to use existing knowledge efficiently has become a problem that companies urgently need to solve.The case is an effective form of implicit knowledge,which has been widely studied in knowledge management.Among them,the most widely used is the case-based reasoning system.The self-learning system can manage cases,and its key link is case retrieval.How to improve the efficiency of case retrieval is the main purpose of this study.For this reason,this thesis improves the fuzzy C-means algorithm and applies it to the case retrieval process.Through the clustering of existing cases,the search efficiency can be improved.The main content of this thesis is how to improve the FCM algorithm and cluster the existing cases.Firstly,the research background of the clustering of implicit knowledge explicit cases based on the improved FCM algorithm is expounded,and the research significance is derived from both theoretical significance and application value.The research status at home and abroad was summarized and analyzed to find out the existing problems.Integrate the problems of the traditional FCM algorithm,find the entry point of the research,and conceive the main research ideas of this thesis.The organization structure of the thesis and the content of each chapter were roughly sorted out.Secondly,the related theories of tacit knowledge,cluster analysis,set theory and fuzzy clustering are introduced.Including: The concept,acquisition and expression of tacit knowledge.The background,meaning and mathematical model of the cluster analysis.Through the introduction of fuzzy set theory,the significance of fuzzy clustering research and the mathematical model of fuzzy clustering are derived.There are also FCM algorithm's objective function,solution process and iterative process.Then,to solve the deficiencies of the FCM algorithm.Firstly,we introduce the degree of similarity and overall similarity that represent the similarity between the cases to replace the Euclidean distance in the FCM algorithm.Then,the density function is introduced and the initial cluster center is found through an iterative process.The target function is weighted with the normalized density value.Finally,the crosshorizon algorithm of the population optimization algorithm is applied to the optimization of the clustering center.Through the continuous iteration of the clustering center,the clustering problem of the existing cases in the CBR system is solved.Finally,the improved FCM algorithm is applied to the customer segmentation of xx logistics company,and according to the clustering results,the logistics company is pointed out the key point of customer relationship management.The example analysis process further validates the effectiveness and operability of the improved FCM algorithm designed and proposed in this thesis.