Application of interpretable machine learning models for the intelligent decision

doi:10.1016/j.neucom.2018.12.012

CSpace > 国家数学与交叉科学中心

	Application of interpretable machine learning models for the intelligent decision
	Li, Yawen 1; Yang, Liu 2; Yang, Bohan 3; Wang, Ning 4; Wu, Tian5,6,7
	2019-03-14
发表期刊	NEUROCOMPUTING
ISSN	0925-2312
卷号	333 页码:273-283
摘要	In this study, an interpretable machine learning algorithm is proposed for the issues of intelligent decision through predicting the firms' efficiency of innovation. Based on the unbalanced panel data collected in Zhongguancun Science Parks from year 2005 to 2015, the efficiency of over 10,000 firms have been analysed in this study, and the change and growth of these firms have been captured over time. The linear regression, decision tree, random forests, neural network and XGBoost models are applied to figure out the impact factors of innovation. After comparing the results of different models, it has been found that the accuracy of XGBoost for R&D efficiency labelled, commercial efficiency labelled and overall efficiency labelled classification problems are 73.65%, 70.02% and 70.09%, which outperform the other four models. Moreover, the interpretability of XGBoost is also better than other models. Thus, the XGBoost model makes it possible for managers to predict the firm's future innovation performance derived from their innovation strategies in the current stage. Furthermore, it helps firms to build an intelligent decision support system, which is of great importance for them to deal with complex decision environments, and to increase their efficiency of innovation in the long-term dynamic competition with other firms. (C) 2018 Elsevier B.V. All rights reserved.
关键词	Machine learning XGBoost R&D investments Firm size Innovation performance
DOI	10.1016/j.neucom.2018.12.012
语种	英语
资助项目	National Natural Science Foundation of China[71804181] ; National Center for Mathematics and Interdisciplinary Sciences, CAS
WOS研究方向	Computer Science
WOS类目	Computer Science, Artificial Intelligence
WOS记录号	WOS:000456834100025
出版者	ELSEVIER SCIENCE BV
引用统计
文献类型	期刊论文
条目标识符	http://ir.amss.ac.cn/handle/2S8OKBNM/32292
专题	国家数学与交叉科学中心
通讯作者	Wu, Tian
作者单位	1.Beijing Univ Posts & Telecommun, Sch Econ & Management, Beijing, Peoples R China 2.Hong Kong Univ Sci & Technol, Dept Comp Sci & Engn, Hong Kong, Peoples R China 3.Chinese Acad Social Sci, Grad Sch, Dept World Econ & Polit, Beijing, Peoples R China 4.China Agr Univ, Int Coll Beijing, Beijing, Peoples R China 5.Chinese Acad Sci, Acad Math & Syst Sci, NCMIS, Beijing, Peoples R China 6.Univ Chinese Acad Sci, Sch Econ & Management, Beijing, Peoples R China 7.Chinese Acad Sci, Key Lab Big Data Min & Knowledge Management, Beijing, Peoples R China
推荐引用方式 GB/T 7714	Li, Yawen,Yang, Liu,Yang, Bohan,et al. Application of interpretable machine learning models for the intelligent decision[J]. NEUROCOMPUTING,2019,333:273-283.
APA	Li, Yawen,Yang, Liu,Yang, Bohan,Wang, Ning,&Wu, Tian.(2019).Application of interpretable machine learning models for the intelligent decision.NEUROCOMPUTING,333,273-283.
MLA	Li, Yawen,et al."Application of interpretable machine learning models for the intelligent decision".NEUROCOMPUTING 333(2019):273-283.