The art of deep learning and natural language processing for emotional sentiment analysis on the academic scholars' peer review process.
Pascal Muam Mah · 2023
Deep learning and natural language processing has emerged as modern machine with state-of-the-art modern learning and applicable techniques that help academicians, students, and teachers to identify, evaluate and validate documents and text features. Most educational systems have incorporated deep learning and natural language processing in fulfilling tasks except publishers and editorial boards, whom they handle important aspects of academic achievement in the decision-making for both journals and conferences. The aim is to introduce a standardized systematic peer review process based on factual justification, amendment, and recommendation approach on scholarly works free from emotional sentiments, politicization, bias, criticism, and condemnation. The study uses peer review feedback from journal XXXX to identify, evaluate and validate emotional sentimental aspects. A Confusion Metrix model was examined in the study to justify and validate the emotional sentiment pros and cons of the reviewer. Also, behavior-oriented drive and influential functions were used to convert text to value and rank the sentiment score. Each reviewer's text was converted into value using a behavior-oriented drive and influential function were used to examine the emotional sentiment level of the reviewers. Text mining and extraction of data uses a natural language process approach to identify sentiments from the reviewer's comments. Based on the confusion matrix and behavior-oriented drive, and influential function, the findings revealed limited significant emotional sentiment involved during the peer review process. The study concluded that the reviewers didn't perfectly validate the manuscript content due to minor emotional sentiments. The reviewer's evaluation of paper #Journal XXXX based on the findings exercise, few emotional sentiments were attracted during the review process. The study recommended editorial boards, journals, conferences, and publishing houses implement this approach to avoid emotional sentiments, politicization, and bias impact on academic evaluation, and validation.