Application of Data Science Technology on Research of Circulatory System Disease Prediction Based on a Prospective Cohort

• Main Authors: Haijing Tang, Guo Chen, Yu Kang, Xu Yang
• Format: Article
• Language: English
• Published: MDPI AG 2018-10-01
• Series: Algorithms
• Subjects: population cohort; circulatory system disease; Cox proportional hazards model; support vector machine; disease prediction
• Online Access: http://www.mdpi.com/1999-4893/11/10/162

Chronic diseases represented by circulatory diseases have gradually become the main types of diseases affecting the health of our population. Establishing a circulatory system disease prediction model to predict the occurrence of diseases and controlling them is of great significance to the health of our population. This article is based on the prospective population cohort data of chronic diseases in China, based on the existing medical cohort studies, the Kaplan–Meier method was used for feature selection, and the traditional medical analysis model represented by the Cox proportional hazards model was used and introduced. Support vector machine research methods in machine learning establish circulatory system disease prediction models. This paper also attempts to introduce the proportion of the explanation variation (PEV) and the shrinkage factor to improve the Cox proportional hazards model; and the use of Particle Swarm Optimization (PSO) algorithm to optimize the parameters of SVM model. Finally, the experimental verification of the above prediction models is carried out. This paper uses the model training time, Accuracy rate(ACC), the area under curve (AUC)of the Receiver Operator Characteristic curve (ROC) and other forecasting indicators. The experimental results show that the PSO-SVM-CSDPC disease prediction model and the S-Cox-CSDPC circulation system disease prediction model have the advantages of fast model solving speed, accurate prediction results and strong generalization ability, which are helpful for the intervention and control of chronic diseases.