Resource Summary Report  New Search Previous Search Results

URL: https://dx.doi.org/10.17504/protocols.io.rv2d68e

Authors: Chu-Yu Chin, Sun-Yuan Hsieh, Vincent S. Tseng

Summary: The early assessment of disease risk is an emerging topic in medical informatics. If diseases are detected at an early stage, prognosis can be improved and medical resources can be used more efficiently. A number of recent studies have considered risk factor analysis approaches, such as association rule mining, sequential rule mining, regression, and medical expert advice. In this study, for improving disease risk assessment, non-negative matrix factorization and support vector machine (SVM) were integrated to discover important and implicit risk factors.To make the method easy to follow, here we provide an experimental protocal. This experimental protocal comprises three main stages: data preprocessing, risk factor optimization, and early disease risk assessment. To discover the optimized risk factors, the NMF algorithm with parameter optimization was used for constructing the NMF-based matrix. In the assessment model learning and early disease risk assessment stages, the machine learning classifier SVM was used for disease modeling with the NMF-based matrix, yielding the final disease risk assessment, which serves as an excellent reference for physicians and patients.

Associated Publications: Chin C, Hsieh S, Tseng VS (2018) eDRAM: Effective early disease risk assessment with matrix factorization on a large-scale medical database: A case study on rheumatoid arthritis. PLoS ONE 13(11): e0207579. doi: 10.1371/journal.pone.0207579

Affiliations: National Cheng Kung University, National Cheng Kung University, National Chiao Tung University

External URL: https://doi.org/10.1371/journal.pone.0207579

Version: 1

Publication Date: 2018

Expand All Collapse