| 학회 |
한국공업화학회 |
| 학술대회 |
2022년 봄 (05/11 ~ 05/13, 제주국제컨벤션센터(ICC JEJU)) |
| 권호 |
26권 1호 |
| 발표분야 |
포스터-촉매 |
| 제목 |
Construction of hierarchical regression model for predicting catalytic performance of non-oxidative propane dehydrogenation |
| 초록 |
Machinelearning techniques have been considered as a powerful tool to predict catalyticperformance and identify influential factors in development of heterogeneous catalysts.In this study, we conducted to hierarchical regression based on sub-datasetsdivided by clustering algorithm in order to construct highly-predictive modelof catalytic performance for non-oxidative propane dehydrogenation. Databasewas composed of 265 points collected from catalytic tests with 160 supported catalysts.Principal component analysis was applied to find out whether the demarcated clusterswere obtained according to catalyst compositions and reaction conditions, and thehierarchical regression was performed using optimally partitioned clusters as sub-datasets.By comparing the prediction accuracy of the hierarchical and conventional models,it was confirmed that the hierarchical regression is an effective strategy to predictcatalytic performance. |
| 저자 |
박지수, 오정목, 전남기, 윤용주
|
| 소속 |
포항공과대 |
| 키워드 |
Machine learning; Propane dehydrogenation
|
| E-Mail |
|
