Machine learning assisted prediction of disperse dye exhaustion on polylactic acid fiber with interpretable model

IF 4.1 3区 工程技术 Q2 CHEMISTRY, APPLIED
Shicheng Liu , Du Chen , Fengxuan Zhang , Qiangqiang Zhao , Jinxin He , Xia Dong
{"title":"Machine learning assisted prediction of disperse dye exhaustion on polylactic acid fiber with interpretable model","authors":"Shicheng Liu ,&nbsp;Du Chen ,&nbsp;Fengxuan Zhang ,&nbsp;Qiangqiang Zhao ,&nbsp;Jinxin He ,&nbsp;Xia Dong","doi":"10.1016/j.dyepig.2025.112693","DOIUrl":null,"url":null,"abstract":"<div><div>Polylactic acid (PLA) is a promising green alternative for petroleum-based synthetic fibers, but the high-exhaustion dyeing of PLA is still an obstacle to its widespread application in textiles and therefore the development of disperse dye for PLA dyeing has been an urgent focus. Here, the exhaustion database of disperse dyes for PLA fiber from literatures and laboratory experiments was established to develop a machine learning model for predicting the dye exhaustion on PLA fiber, and the model was interpreted by Shapley Additive exPlanations (SHAP) and applied to pre-filtering out candidates with high-exhaustion that collected from literatures. It was found that the AUC of the constructed model in 10-fold stratified cross-validation and test set were 0.887 and 0.859, respectively. According to SHAP analysis, chain substructures such as the ester chain and alkyl chain are conducive to exhaustion while the cyan group (∗-C<img>N) attached to the aromatic ring is unfavorable. In external application, the model maintained an AUC of 0.885, demonstrating excellent applicability and generalizability. Furthermore, 3 yellow dyes from 27 reported samples were screened out as worthy of high-exhaustion dyeing because of the shortage of yellow dyes for PLA fiber. This study provides a convenient way to develop high-performance dyes for new green fibers. All data and code are available from Github (<span><span>https://github.com/Sixty-four-floor/Exhaustion-PLA</span><svg><path></path></svg></span>).</div></div>","PeriodicalId":302,"journal":{"name":"Dyes and Pigments","volume":"237 ","pages":"Article 112693"},"PeriodicalIF":4.1000,"publicationDate":"2025-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Dyes and Pigments","FirstCategoryId":"88","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0143720825000634","RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"CHEMISTRY, APPLIED","Score":null,"Total":0}
引用次数: 0

Abstract

Polylactic acid (PLA) is a promising green alternative for petroleum-based synthetic fibers, but the high-exhaustion dyeing of PLA is still an obstacle to its widespread application in textiles and therefore the development of disperse dye for PLA dyeing has been an urgent focus. Here, the exhaustion database of disperse dyes for PLA fiber from literatures and laboratory experiments was established to develop a machine learning model for predicting the dye exhaustion on PLA fiber, and the model was interpreted by Shapley Additive exPlanations (SHAP) and applied to pre-filtering out candidates with high-exhaustion that collected from literatures. It was found that the AUC of the constructed model in 10-fold stratified cross-validation and test set were 0.887 and 0.859, respectively. According to SHAP analysis, chain substructures such as the ester chain and alkyl chain are conducive to exhaustion while the cyan group (∗-CN) attached to the aromatic ring is unfavorable. In external application, the model maintained an AUC of 0.885, demonstrating excellent applicability and generalizability. Furthermore, 3 yellow dyes from 27 reported samples were screened out as worthy of high-exhaustion dyeing because of the shortage of yellow dyes for PLA fiber. This study provides a convenient way to develop high-performance dyes for new green fibers. All data and code are available from Github (https://github.com/Sixty-four-floor/Exhaustion-PLA).
求助全文
约1分钟内获得全文 求助全文
来源期刊
Dyes and Pigments
Dyes and Pigments 工程技术-材料科学:纺织
CiteScore
8.20
自引率
13.30%
发文量
933
审稿时长
33 days
期刊介绍: Dyes and Pigments covers the scientific and technical aspects of the chemistry and physics of dyes, pigments and their intermediates. Emphasis is placed on the properties of the colouring matters themselves rather than on their applications or the system in which they may be applied. Thus the journal accepts research and review papers on the synthesis of dyes, pigments and intermediates, their physical or chemical properties, e.g. spectroscopic, surface, solution or solid state characteristics, the physical aspects of their preparation, e.g. precipitation, nucleation and growth, crystal formation, liquid crystalline characteristics, their photochemical, ecological or biological properties and the relationship between colour and chemical constitution. However, papers are considered which deal with the more fundamental aspects of colourant application and of the interactions of colourants with substrates or media. The journal will interest a wide variety of workers in a range of disciplines whose work involves dyes, pigments and their intermediates, and provides a platform for investigators with common interests but diverse fields of activity such as cosmetics, reprographics, dye and pigment synthesis, medical research, polymers, etc.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信