Faster Case Retrieval Using Hash Indexing Technique

The main objective of case retrieval is to scan and to map the most similar old cases in case basewith a new problem. Beside accurateness, the time taken to retrieve case is also important. Withthe increasing number of cases in case base, the retrieval task is becoming more challengingwhere faster retrieval time and good accuracy are the main aim. Traditionally, sequential indexingmethod has been applied to search for possible cases in case base. This technique worked fastwhen the number of cases is small but requires more time to retrieve when the number of data incase base grows. As an alternative, this paper presents the integration of hashing indexingtechnique in case retrieval to mine large cases and speed up the retrieval time. Hashing indexingsearches a record by determining the index using only an entry’s search key without traversing allrecords. To test the proposed method, real data namely Timah Tasoh Dam operational dataset,which is temporal in nature that represents the historical hydrological data of daily Timah Tasohdam operation in Perlis, Malaysia ranging from year 1997-2005, was chosen as experiment.Then, the hashing indexing performance is compared with sequential method in term of retrievaltime and accuracy. The finding indicates that hashing indexing is more accurate and faster thansequential approach in retrieving cases. Besides that, the combination of hashing search keyproduces better result compared to single search key.