Fuzzy cognitive map approach to web-mining inference amplification

Abstract

Abstract This paper is concerned with proposing the fuzzy cognitive map (FCM)-driven inference amplification mechanism in the field of web-mining. As the recent advent of the Internet, most of the modern firms are now geared towards using the web technology in their daily as well as strategic activities. The web-mining technology provides them with unprecedented ability to analyze web-log data, which are seemingly full of useful information, but often lack of important and meaningful information. This indicates the need to develop an advanced inference mechanism extracting richer implication from the web-mining results. In this sense, we propose a new web-mining inference amplification (WEMIA) mechanism using the inference logic of FCM. The association rule mining is what we adopt as the web-mining technique to prove the validity of the proposed WEMIA. The main recipe of the proposed WEMIA is the three-phased inference amplification. The first phase is to apply the association rule mining, and the second phase is to transform the association rules into FCM-driven causal knowledge bases. The third phase is dedicated to amplifying the inference by developing the causal knowledge-based inference equivalence property, which was derived from analyzing the inference mechanism of FCMs. With an illustrative web-log database, we suggest results proving the robustness of our proposed WEMIA mechanism.