Analysis of Biomedical Entities extraction and classification using TensorFlow based on Neural Network

Abstract

Deep Learning is an important research and nowadays growing area. Deep learning does not require any process tasks or it require simply feature generation process instead; techniques are used for automatically extracted the features. The major biological applications are included in many research areas like biomarker development, drug discovery and prediction, gene annotation, medical image recognition, repurposing and structural biology, chemistry. Over the past decades biomedical entities classification have been focused by many research areas for improving the entities identification and classification accuracy, but still the problem persists due to the nomenclature and the new entities are identified day by day. The paper concentrated on the biomedical entities classification. There are different ways of entities identification approaches based on the entities. The methodology includes four phases: the first phase data collection and the second phase text data preprocessing; text are preprocessed using various text preprocessing library like NLTK; the processed data are classified using TensorFlow in neural network. An experimental study was done on various text based sentences which are collected from NCBI database. The overall performance of the proposed method is examined in terms of accuracy. The method exhibited 87% extracting entities classification which shows the significant performance.