Statistical Analysis of Web Big Data Based on Dynamic Backward Regression Equation

Abstract

While there are various devices and strategies for different regression, there are not many for stepwise numerous regressions. The ongoing form of BACKSTEP utilizes a backward factor choice technique that starts with an assortment of n factors and, in view of a measurable convenience standard, specifically erases each factor in turn to make continuously more modest subsets of indicators. Despite the fact that it is more uncommon than forward determination, backward choice produces programs that are fairly straightforward. As the economy has developed, brilliant networks stand out and utilize. Be that as it may, there are as yet various issues with the framework in this field, including unreasonable structure costs, an absence of knowledge, the shared freedom of numerous frameworks, the test of bound together administration, and others. In this period of state-of-the-art innovation, data is quick growing to frame very immense data assortments, or "Big Data," all around the locale. The examination of immense data to get wrathful data from them is quite possibly of the most significant and testing task in data analytics research. The main contributing factors to how traditional data analysis on big data is seen are the difficulties of limited memory use, reduced processing speed, and computational barrier. We give a thorough analysis of regression analysis on a predictive big data model in this work. The regression model is a crucial tool for data modelling and analysis. The suggested model contains three phases in this chapter. The ELR model's performance is examined empirically using a sizable diabetes dataset. The findings demonstrate that the ELR model is more accurate than the conventional logistic regression approach.