Quantitative Analysis Powered by Naïve Bayes Classifier Algorithm to Data-Related Publications Social-Scientific Network

Quantitative evaluation of a dataset can play an important role in pattern recognition of technical-scientific research involving behavior and dynamics in social networks. As an example, are the adaptive feature weighting approaches by naive Bayes text algorithm. This work aims to present an exploratory data analysis with a quantitative approach that involves pattern recognition using the Mendeley research network; to identify logics given the popularity of document access. To better analyze the results, the work was divided into four categories, each with three subcategories, that is, five, three, and two output classes. The name for these categories came up due to data collection, which also presented documents with open access, dismembering proceedings, and journals for two more categories. As a result, the performance for the test examples showed a lower error rate related to the subcategory two output classes in the criterion of popularity by using the naive Bayes algorithm in Mendeley