SVM-based knowledge topic identification toward the autonomous knowledge acquisition

One of the most serious problems that conventional knowledge management (KM) encompasses has been pointed out tardy and ineffective acquisition of knowledge. To resolve this problem, knowledge must be autonomously acquired according to its context of use by applying the technique of keyword extraction in machine learning algorithm-based text mining. Once the topic of the given knowledge can be identified in an automated manner, then a set of knowledge can be explicitly stored in a knowledge repository; fully automated acquisition of knowledge can be achieved. This paper, therefore, suggests an amended knowledge acquisition framework, especially focused on the autonomous acquisition of knowledge in ordinary dialogues. The suggested methodology is underpinned by the functionality of the support vector machine (SVM) which was demonstrated to identify the topic of knowledge in the most accurate and efficient way. To validate the feasibility of the proposed concepts, CKAS (Context-based Knowledge Acquisition System), a prototype system, is implemented.