Dynamic Recommendation: Disease Prediction and Prevention Using Recommender System

International Journal of Basic Science in Medicine

Amir Kiani¹  Mahdi Nasiri²  Behrouz Minaei³ 

[1] ;Computer Engineering Department, Iran University of Science and Technology (IUST), Tehran, Iran;Department of Mathematics and Computer Science, Amir Kabir University of Technology, Tehran, Iran;

关键词: Recommender system;    Disease prediction;    Collaborative filtering;    Data mining;    Treatment;   

DOI  :  10.15171/ijbsm.2016.04

来源: DOAJ

【 摘 要 】

Background: In today’s world, chronic diseases are predominant health problems and cause heavy burden on society; therefore early diagnosis and even prediction of the disease is a way to reduce this burden. In this project, we tried to use recommender system to predict which other diseases a chronic patient is susceptible for. Methods: In this study, through a dynamic recommender system, we evaluated patients’ treatment destiny during the time. Results: It was shown that our method increased accuracy and reduced error compared with other recommendation methods in disease prediction. Conclusion: Compared to current usual methods, in our method we used previous patients’ characteristics as one of the factorization variables to predict destiny of future patients. Furthermore, using this method, we can predict which complication or disease the patient would suffer from first in future. Therefore, we can manage policies toward disease burden reduction by implementing prevention programs.

【 授权许可】

Unknown