{"title":"电厂小样本建模的归纳转移回归框架","authors":"","doi":"10.1016/j.cherd.2024.08.020","DOIUrl":null,"url":null,"abstract":"<div><p>Small sample size presents a significant challenge in process modeling, making machine learning (ML) models prone to overfitting and reduced accuracy. To address this issue, this study develops a novel inductive transfer regression framework called double-weight least squares support vector regression (DWLSSVR). First, sample weights are incorporated to minimize the multi-kernel maximum mean discrepancy (MK-MMD) between domains, thereby promoting joint distribution adaptation and decreasing domain discrepancy. Second, the impact of unrelated source domain samples is further mitigated by iterative weights derived from fitting errors. In addition, a two-step strategy is developed to optimize the hyperparameters in DWLSSVR, which introduces a new criterion based on Wasserstein distance (WD). A numerical simulation demonstrates the effectiveness of the developed framework. Then, the proposed method is applied to the small sample modeling of a complex chemical process. The results of predicting NO<sub><em>x</em></sub> emissions from a coal-fired boiler demonstrate that the DWLSSVR model achieves superior prediction accuracy, with a coefficient of determination (R<sup>2</sup>) of 0.942 under the new operating condition. In contrast, the best LSSVR model achieves an R<sup>2</sup> of 0.844 under the same condition.</p></div>","PeriodicalId":10019,"journal":{"name":"Chemical Engineering Research & Design","volume":null,"pages":null},"PeriodicalIF":3.7000,"publicationDate":"2024-08-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"An inductive transfer regression framework for small sample modeling in power plants\",\"authors\":\"\",\"doi\":\"10.1016/j.cherd.2024.08.020\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><p>Small sample size presents a significant challenge in process modeling, making machine learning (ML) models prone to overfitting and reduced accuracy. To address this issue, this study develops a novel inductive transfer regression framework called double-weight least squares support vector regression (DWLSSVR). First, sample weights are incorporated to minimize the multi-kernel maximum mean discrepancy (MK-MMD) between domains, thereby promoting joint distribution adaptation and decreasing domain discrepancy. Second, the impact of unrelated source domain samples is further mitigated by iterative weights derived from fitting errors. In addition, a two-step strategy is developed to optimize the hyperparameters in DWLSSVR, which introduces a new criterion based on Wasserstein distance (WD). A numerical simulation demonstrates the effectiveness of the developed framework. Then, the proposed method is applied to the small sample modeling of a complex chemical process. The results of predicting NO<sub><em>x</em></sub> emissions from a coal-fired boiler demonstrate that the DWLSSVR model achieves superior prediction accuracy, with a coefficient of determination (R<sup>2</sup>) of 0.942 under the new operating condition. In contrast, the best LSSVR model achieves an R<sup>2</sup> of 0.844 under the same condition.</p></div>\",\"PeriodicalId\":10019,\"journal\":{\"name\":\"Chemical Engineering Research & Design\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":3.7000,\"publicationDate\":\"2024-08-24\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Chemical Engineering Research & Design\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S0263876224004982\",\"RegionNum\":3,\"RegionCategory\":\"工程技术\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"ENGINEERING, CHEMICAL\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Chemical Engineering Research & Design","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0263876224004982","RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ENGINEERING, CHEMICAL","Score":null,"Total":0}
An inductive transfer regression framework for small sample modeling in power plants
Small sample size presents a significant challenge in process modeling, making machine learning (ML) models prone to overfitting and reduced accuracy. To address this issue, this study develops a novel inductive transfer regression framework called double-weight least squares support vector regression (DWLSSVR). First, sample weights are incorporated to minimize the multi-kernel maximum mean discrepancy (MK-MMD) between domains, thereby promoting joint distribution adaptation and decreasing domain discrepancy. Second, the impact of unrelated source domain samples is further mitigated by iterative weights derived from fitting errors. In addition, a two-step strategy is developed to optimize the hyperparameters in DWLSSVR, which introduces a new criterion based on Wasserstein distance (WD). A numerical simulation demonstrates the effectiveness of the developed framework. Then, the proposed method is applied to the small sample modeling of a complex chemical process. The results of predicting NOx emissions from a coal-fired boiler demonstrate that the DWLSSVR model achieves superior prediction accuracy, with a coefficient of determination (R2) of 0.942 under the new operating condition. In contrast, the best LSSVR model achieves an R2 of 0.844 under the same condition.
期刊介绍:
ChERD aims to be the principal international journal for publication of high quality, original papers in chemical engineering.
Papers showing how research results can be used in chemical engineering design, and accounts of experimental or theoretical research work bringing new perspectives to established principles, highlighting unsolved problems or indicating directions for future research, are particularly welcome. Contributions that deal with new developments in plant or processes and that can be given quantitative expression are encouraged. The journal is especially interested in papers that extend the boundaries of traditional chemical engineering.