ABSTRACT

Due to the tremendous growth of data and large databases, efficient extraction of required data has become a challenging task. This paper propose a novel approach for knowledge discovery from huge unlabeled temporal databases by employing a combination of HMM and K-means technique. We propose to recursively divide the entire database into clusters having similar characteristics, this process is repeated until we get the cluster’s where no further diversification is possible. Thereafter, the clusters are labeled for knowledge extraction for various purposes.