Max-margin weight learning for medical knowledge network. (March 2018)

• Record Type: Journal Article
• Authors: Jiang, Jingchi; Xie, Jing; Zhao, Chao; Su, Jia; Guan, Yi; Yu, Qiubin
• Is Part Of: Computer methods and programs in biomedicine. Volume 156(2018)
• Page Start: 179
• Page End: 190
• Publication Date: 2018-03
• Subjects: Markov logic network -- Medical knowledge network -- Weight learning -- Electronic medical records; Medicine -- Computer programs -- Periodicals; Biology -- Computer programs -- Periodicals; Computers -- Periodicals; Medicine -- Periodicals; Médecine -- Logiciels -- Périodiques; Biologie -- Logiciels -- Périodiques; Biology -- Computer programs; Medicine -- Computer programs; Periodicals; Electronic journals; 610.28
• DOI: 10.1016/j.cmpb.2018.01.005
• Languages: English
• ISSNs: 0169-2607

Abstract

Highlights: Topological structure of MKN for representing first-order knowledge. Probabilistic graphical models used for medical diagnosis. Discriminative weight learning method of MKN based on a max-margin framework. Proposed approach is superior to classical machine learning for medical diagnosis. Abstract: Background and objective: The application of medical knowledge strongly affects the performance of intelligent diagnosis, and method of learning the weights of medical knowledge plays a substantial role in probabilistic graphical models (PGMs). The purpose of this study is to investigate a discriminative weight-learning method based on a medical knowledge network (MKN). Methods: We propose a training model called the maximum margin medical knowledge network (M 3 KN), which is strictly derived for calculating the weight of medical knowledge. Using the definition of a reasonable margin, the weight learning can be transformed into a margin optimization problem. To solve the optimization problem, we adopt a sequential minimal optimization (SMO) algorithm and the clique property of a Markov network. Ultimately, M 3 KN not only incorporates the inference ability of PGMs but also deals with high-dimensional logic knowledge. Results: The experimental results indicate that M 3 KN obtains a higher F-measure score than the maximum likelihood learning algorithm of MKN for both Chinese Electronic Medical Records (CEMRs) and Blood Examination Records (BERs). Furthermore, the proposed approach is obviously superior to some classical machine learning algorithms for medical diagnosis. To adequately manifest the importance of domain knowledge, we numerically verify that the diagnostic accuracy of M 3 KN is gradually improved as the number of learned CEMRs increase, which contain important medical knowledge. Conclusions: Our experimental results show that the proposed method performs reliably for learning the weights of medical knowledge. M 3 KN outperforms other existing methods by achieving an F -measure of 0.731 for CEMRs and 0.4538 for BERs. This further illustrates that M 3 KN can facilitate the investigations of intelligent healthcare.