An Intelligent Smart Ranked Feature Construction Analysis based on Clustering High Dimensional Data Streams

Abstract

Artificial Intelligence is nowadays successfully applied to massive data sets assembled from various areas. One of the noteworthy challenges in applying AI methodologies to massive data sets is the way by which to effectively use accessible computational resources when building prescient and inferential models, while utilizing data in a measurably optimal manner. Random projections have been dynamically received for a differing set of tasks in AI including dimensional decrease. One explicit line of research on this point has analyzed the use of quantization rebutting in projection with the point of additional data pressure. We present a fundamental calculation, named double random projection, which uses the double solution of the low-dimensional improvement issue to recuperate the ideal solution for the first issue.. Our Hypothetical Examination (HE) shows that with a high probability, the proposed calculation can accurately recoup the ideal solution for the first issue, given that the data lattice is (roughly) low-position and ideal solution is (around) inadequate.. We further exhibit that the proposed calculation can be applied iteratively to diminishing the exponentially