Identification of Diseases in Clinical Support System Using Extended Graph Convolution Neural Networks
European Journal of Molecular & Clinical Medicine,
2020, Volume 7, Issue 10, Pages 3374-3384
Abstract
Disease Identification, using Patient Symptoms is a important classical problem in clinical industry. Disease Identification using Natural Language Processing (NLP) requires efficient method for classifying medical data. Number of studies that applies convolution neural network for classification exists and some few flexible methods explored for graph convolution network for NLP based (text based) data classification . In this paper we proposed a method for NLP(text) based graph convolution neural network works both for document and record based structure. In this, labeled data are classified using supervised method and unlabeled data are classified using unsupervised method. We are using corpus based word co-occurrence and word relation processing, then learning process carry out by our proposed text GCN then both labeled and unlabeled data processed as per supervised and unsupervised learning. Text GCN also Reduce percentage of training data. This will increase the performance and robustness of GCN based text classification and prediction method.
Keywords:
- Article View: 97
- PDF Download: 144