Discovering Concealed Semantics in Web Documents Using Fuzzy Clustering By Feature Matrix Methodology

As the data grows exponentially exploding on the 'World Wide Web', the orthodox clustering algorithms obligate various challenges to tackle, of which the most often faced challenge is the uncertainty. Web documents have become heterogeneous and very complex. There exist multiple relations between one web document and others in the form of entrenched links. This can be imagined as a one to many (1-M) relationship, for example a particular web document may fit in many cross domains viz. politics, sports, utilities, technology, music, weather forecasting, linked to e-commerce products etc. Therefore there is a necessity for efficient, effective and constructive context driven clustering methods. Orthodox or the already well-established clustering algorithms adhere to classify the given data sets as exclusive clusters. Signifies that we can clearly state whether to which cluster an object belongs to. But such a partition is not sufficient for representing in the real time. So, a fuzzy clustering method is presented to build clusters with indeterminate limits and allows that one object belongs to overlying clusters with some membership degree. In supplementary words, the crux of fuzzy clustering is to contemplate the fitting status to the clusters, as well as to cogitate to what degree the object belongs to the cluster.