A Fuzzy Inference Approach for Association Rule Mining

Abstract

The association rule mining is most popular and real time applicable approach for finding interesting relations between items. Many of the ARM (Association rule Mining) approaches are well investigated in the literature, but it generates large number of association rules. If the dataset size is larger, then huge rules may occur, often it is a critical situation where decision making is difficult or unattainable because knowledge is not directly present in frequent patterns. This paper presents an improved AIRM (Association Inference Rule Mining) algorithm where fuzzy logic based C-Means clustering concept has been adopted to discover inference knowledge from frequent patterns. For experimental study, we apply this approach on a clinical dataset of 1000 patients, contained symptoms having different diseases. Proposed approach follows three phase procedure in order to achieve inference knowledge, in the first phase preprocess the data, second phase apply the ARM and finally FIS has to be applied to discover inference knowledge by matching inference rules and put the data in the appropriate class on the basis of their matching degree. The new approach is efficient and outperforms as compared to a previous AIRM algorithm in order to match inference rules and knowledge discovery process.

Stock Market Trend Prediction Model Using Data Mining Techniques

“…It became more popular due to its high applicability in various data analysis fields such as DNA pattern recognition, web data mining, clinical data mining, software bug analysis and stock market analysis. Apriori and FP-Growth algorithms are two most frequent pattern mining algorithms in literature [6]. Apriori algorithm which was introduced by Agarwal and Srikant [7] adopts the apriori property and candidate generation process to generate association rules.…”