커뮤니티
교수동정
- 국립 부경대학교 교수님들의 자랑스러운 소식입니다.
| PKNU students run top | Kim Joon-cheol, Lee Seung-jae, and Ko Nak-gyeong | |||
| 작성자 | Department of External Cooperation | 작성일 | 2020-11-23 |
| 조회수 | 297 | ||
| PKNU students run top | Kim Joon-cheol, Lee Seung-jae, and Ko Nak-gyeong | |||||
 |
Department of External Cooperation |  |
2020-11-23 |  |
297 |
|
Students of PKNU swept the thesis presentation awards of the Korean Data & Information Science Society At the recent academic conference held at Keimyung University, Lee Seung-jae from the Department of Artificial Intelligence Convergence (2nd-year master) and Ko Nak-gyeong (2nd-year master) each have won the excellence award and the encouragement award at the thesis presentation award including Kim Joon-cheol (1st-year master) of Statistics received the 1st prize for thesis presentation. Kim Joon-cheol, who received the Grand Prize, was recognized for his outstanding research achievements for his thesis ’M-spline approach for mixture cure survival models (Advisor Ha Il-do)’. In this paper, unlike traditional survival data, certain patients do not experience an event of interest even after sufficient time, so the corresponding survival function converges to a nonzero value (which is called a cure rate), and he proposes a mixture cure models that can be applied to the M-spline approach the above survival data, and received the best evaluation by revealing its validity through an illustration using actual clinical data. Lee Seung-jae, who received the Excellence Award, evaluated how the traditional method of variable selection or penalized regression affects the model evaluation measures and computer computation speed in deep learning regression and classification through real data. His thesis, ’A study on the influence of variable selection in the deep learning regression/classification (advisory prof. Jang Dae-heung)’ was recognized for his achievements. Ko Nak-gyeong, who received the Encouragement Award, proposed an accelerated risk model using log-logistic and generalized extreme value distribution as a method for dealing with clinical trial data in her thesis ’Accelerated Risk Analysis Model for Cross-Risk Data’, and she supported the claim with actual clinical data to prove its validity and received good reviews. Meanwhile, the Korean Data & Information Science Society, which was founded in 1989, is a leading academic activity in the fields related to the 4th industrial revolution, big data, artificial intelligence, and data science. <Pukyong Today> |
