Machine Learning Mechanisms for Big Scholarly Data Analysis

Abstract

This work presents conventional text documents have made it possible to efficiently newlineretrieve large amounts of text data with the development of various search engines. newlineHowever, these traditional search approaches frequently have lower accuracy in newlineretrieval, particularly when documents have certain characteristics that call for more newlinein-depth semantic extraction. A search engine for algorithms called Algorithm Seer newlinehas recently been developed. The normal search engine collects the deep textual newlinemetadata and pseudo-codes from research papers. However, such a system is unable newlineto accommodate user searches that attempt to identify algorithm-specific information, newlinesuch as the datasets on which algorithms operate their effectiveness, runtime newlinecomplication, etc. A number of improvements to the previously suggested algorithm newlinesearch engine are given in this study. Hence this work presents anefficient framework newlinefor Machine Learning Mechanisms for Big Scholarly Data Analysis to develop newlinethis framework this research work is carried out in three different stages referred as newlineobjectives of the research work. newlineFirst objective of the research work is Classifying Big Scholarly data using newlineImproved Naïve Bayes Technique . This work presents Improved Naïve Bayes newlinetechnique to classify and analyze the Scholarly dataset, which shall help the newlineresearches to get the classified data. Naïve Bayes classification algorithm is a very newlinepopular and efficient technique based on probability. It is based on Bayes probability newlinerule, used to compute the probability of an event s occurrence under given conditions newlinealso known as conditional probability. Two assumptions governing Naïve Bayes newlineclassifier are feature independence i.e. each feature is independent of any other in an newlineinput and every feature has equal contribution to theoutput of the classification. The newlineperformance of the proposed technique is assessed by utilizing classification newlineaccuracy, F- measure, Precision and Recall. The proposed Improve NB classifier newlineoutperforms the existing classifiers by 10-1