Original Approach for Reduction of High Dimensionality In unsupervised learning

with the appearance of web 2.0, the huge amount of web services and the increasing number of information, articles, and products placed on line, the quantity of data is exploding throughout the world. In addition, this huge amount of data is qualified as high-dimensional data. Analyzing large datasets is an urgent problem of great practical importance. Precisely, the major concerns are directed to the reduction of high dimensionality of the feature space owing to computational complexity and accuracy consideration. Consequently, variations of methods have been originally introduced in the literature to select an optimal number of features. In this paper, we propose an original method based on a new version of k-means algorithm to reduce the dimension of large data sets thanks to a new proposing process of features' clustering.