Rule induction using Rough Set Theory — An application in agriculture

Rough Set Theory (RST), proposed by Z Pawlak, is a new mathematical approach to vagueness and uncertainty. Tools based on RST are found to be useful in addressing data mining tasks such as classification, clustering and rule mining. In RST all computations are performed directly on the supplied data and works by making use of the granularity structure of the data. Association rules, which play an important role in data mining, provide associations among attributes and generally they are helpful for decision making. A problem of using conventional association rule algorithms is that too many rules are generated by these algorithms and it is very difficult to analyze these rules. This paper proposes a rough set based approach to generate rules from an inconsistent information system consisting of the preprocessed data collected from coconut cultivators of the Keezhur Chavassery Grama Panchayath using stratified random sampling method. An existing algorithm, namely, Learning from Examples Module version 2 (LEM2) is modified to incorporate some conditions, leading to the generation of significant rules. By applying the proposed algorithm, a set of significant rules are generated. These rules are expected to be helpful to the farmers of the state to design their farming plans, which will enable them to improve their coconut production.