شماره ركورد كنفرانس
5255
عنوان مقاله
Depression Diagnosing using Privacy-Preserving Distributed Data Mining
پديدآورندگان
Department of Computer Engineering and IT, Shiraz University of Technology, Iran A., Namdar asrin.namdar@gmail.com , Department of Computer Engineering and IT, Shiraz University of Technology, Iran P., Shamsinejadbabaki p.shamsinejad@sutech.ac.ir , Department of Computer Engineering and IT, Shiraz University of Technology, Iran R., Khayami khayami@sutech.ac.ir
تعداد صفحه
2
كليدواژه
federated learning , secure multiparty computation , Secret Sharing , Federated Pearson Correlation
سال انتشار
1401
عنوان كنفرانس
اولين سمپوزيوم بين المللي كاربردهاي هوش مصنوعي
زبان مدرك
انگليسي
چكيده فارسي
According to WHO statistics 280 million people in the world have depression and the number of people suffering from it increases every year. Timely diagnosis and treatment of this disease can be very important for our society. Despite of all Machine Learning achievements in automatic disease detection, people are not very eager to share their personal data with knowledge-extractor companies because people valorize their private data. In this paper a privacy-preserving depression diagnosing system has been proposed. Our system maintains data confidentiality by using SMPC and federated learning methods. The results of evaluating proposed system on a dataset of depressed patients show that our system has succeeded in diagnosing depression with 93.33% accuracy while maintaining data confidentiality.
كشور
ايران
لينک به اين مدرک