Text Classification for Edentulous and Comorbids Disease Systematic Literature Review based on Machine Learning Algorithm

Systematic Literature Review (SLR) is a secondary research technique used to understand a particular research issue comprehensively. Dentists must do a systematic literature review to keep their knowledge and skills relevant to current patient treatment. The SLR procedure is, nevertheless, relatively time-consuming. This research aims to construct a machine-learning model capable of classifying the language of scientific papers to accelerate the SLR screening procedure—data collection, processing, and selecting of experimental models comprised of the methodology employed in this study. In the case of studies, data from the most recent SLRs are supplemented with particular keyword searches in available papers. Two classes are manually assigned to the manually labeled datasets. The screening of experiments was separated into title-abstract and full-text categories. Multinomial Nave Bayes, Support Vector Machine, Simple Dense, Long-short Term Memory, Gated Recurrent Unit, Bi-directional Long-short Term Memory, Convolutional Neural Network, and Universal Sentence Encoder are the eight classification models. The K-Fold Cross Validation evaluation findings revealed that the SVM model had the highest F1-Score (0.97), followed by the Convolutional Neural Network (0.88) and Multinomial Nave Bayes (0.87).