Ontology-Based Framework for Personalized Diagnosis and Prognosis of Cancer Based on Gene Expression Data

摘要

This paper presents an ontology-based framework for personalized cancer decision support system based on gene expression data. This framework integrates the ontology and personalized cancer predictions using a variety of machine learning models. A case study is proposed for demonstrating the personalized cancer diagnosis and prognosis on two benchmark cancer gene data. Different methods based on global, local and personalized modeling, including Multi Linear Regression (MLR), Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Evolving Classifier Function (ECF), weighted distance weighted variables K-nearest neighbor method (WWKNN) and a transductive neuro-fuzzy inference system with weighted data normalization (TWNFI) are investigated. The development platform is general that can use multimodal information for personalized prediction and new knowledge creation within an evolving ontology framework.