A New Approach for Optimization Classification Rule Generation Technique

Data mining techniques enable an application to analyse a rich amount of data and recover the essential information from it. This information can use for decision making, pattern recognition and other applications. This data model can be a transparent data structure or a set of rules. In this presented work the transparent data models are investigated for optimizing their performance and resource consumption. During experiments that are observed these data models are accurately identify the patterns as defined the classification rules but the number of comparisons for classification increases as the number of rules generated are increases. The proposed technique first analyse the entire data samples and then the most optimum attributes are targeted for rule development. The proposed classification rule generation technique is efficiently generating less number of rules as compared to the traditionally available techniques. The implementation of the proposed concept is provided using MATLAB simulation tool and the performance in terms of memory consumption, time consumption and numbers of rules are evaluated. According to the obtained results the performance of the proposed rule generation technique is much efficient as compared to the traditionally available techniques.