XBNet and Text Mining-based Genetic Diseases Classification

Text mining can be used for various biological and medical purposes. In this work, we have used text mining to analyze various papers related to Alzheimer's, Asthma, Cancer, Diabetes, Fabry, and various syndrome diseases using their abstracts and extracting from the text the related gene to those diseases. In this case, the data set collected from the various sources was generated manually in this pilot study. The text was collected from various articles searched through PubMed, Web of Science, and Medline. Extremely Boosted Neural Network is a new machine learning (ML) algorithm that has been developed recently to train a new optimization technique and can integrate tree-based models with Neural networks (NN). In this paper, an extremely boosted neural network was utilized as a novel application for the analysis of text data to extract information about the diseases. We benchmark the proposed model with 17 other ML models, achieving 98% accuracy. This is a significant improvement given that most of the benchmark techniques received less than 97%.