Techniques for Efficient Case Retrieval and Rainfall Prediction Using CBR and Fuzzy Logic

Case based reasoning means to understand and solve new problems using past experiences. At the first stage this paper reports different phases of case based reasoning, methods and techniques involved in case retrieval phase and how these techniques are used by different researchers to make retrieval phase more significant in retrieving most similar cases. It also outlines the probable scope of the works in case base retrieval. At the second stage, this paper introduces a rainfall prediction method by fuzzy case based reasoning. In our study techniques used to rainfall prediction are investigated and then we have applied hybridization of fuzzy logic and case base reasoning for rainfall prediction. We have treated the important weather attributes of a day as a case. Fuzzy k-NN is applied to retrieve similar cases from case-base and we find the solution of the new case by adapting solutions of retrieved cases. A case-base of 10,000 cases is build. The obtained result shows promising prediction of rainfall.