{"title":"机器学习用于绩效归因陈述的可行性分析","authors":"Anil Berkin , Walter Aerts , Tom Van Caneghem","doi":"10.1016/j.accinf.2022.100597","DOIUrl":null,"url":null,"abstract":"<div><p>We investigate the feasibility of machine learning methods for attributional content and framing analysis in corporate reporting. We test the performance of five widely-used supervised machine learning classifiers (naïve Bayes, logistic regression, support vector machines, random forests, decision trees) in a top-down three-level hierarchical setting to (1) identify performance-related statements; (2) detect attributions in these; and (3) classify the content of the attributional statements. The training set comprises manually coded statements from a corpus of management commentary reports of listed companies. The attributions include both intra- and inter-sentential attributional statements. The results show that for both intra- and inter-sentential attributions, F1-scores of our most accurate classifier (i.e., support vector machines) vary in the range of 76% up to 94%, depending on the identification, detection and classification levels and the content characteristics of attributions. Additionally, we assess the hierarchical performance of classifiers, providing insights into a more holistic classification process for attributional statements. Overall, our results show how machine learning methods may facilitate narrative disclosure analysis by providing a more efficient way to detect and classify performance-related attributional statements. Our findings contribute to the accounting and management literature by providing a basis for implementing machine learning methodologies for research investigating attributional behavior and related impression management.</p></div>","PeriodicalId":47170,"journal":{"name":"International Journal of Accounting Information Systems","volume":"48 ","pages":"Article 100597"},"PeriodicalIF":4.1000,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":"{\"title\":\"Feasibility analysis of machine learning for performance-related attributional statements\",\"authors\":\"Anil Berkin , Walter Aerts , Tom Van Caneghem\",\"doi\":\"10.1016/j.accinf.2022.100597\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><p>We investigate the feasibility of machine learning methods for attributional content and framing analysis in corporate reporting. We test the performance of five widely-used supervised machine learning classifiers (naïve Bayes, logistic regression, support vector machines, random forests, decision trees) in a top-down three-level hierarchical setting to (1) identify performance-related statements; (2) detect attributions in these; and (3) classify the content of the attributional statements. The training set comprises manually coded statements from a corpus of management commentary reports of listed companies. The attributions include both intra- and inter-sentential attributional statements. The results show that for both intra- and inter-sentential attributions, F1-scores of our most accurate classifier (i.e., support vector machines) vary in the range of 76% up to 94%, depending on the identification, detection and classification levels and the content characteristics of attributions. Additionally, we assess the hierarchical performance of classifiers, providing insights into a more holistic classification process for attributional statements. Overall, our results show how machine learning methods may facilitate narrative disclosure analysis by providing a more efficient way to detect and classify performance-related attributional statements. Our findings contribute to the accounting and management literature by providing a basis for implementing machine learning methodologies for research investigating attributional behavior and related impression management.</p></div>\",\"PeriodicalId\":47170,\"journal\":{\"name\":\"International Journal of Accounting Information Systems\",\"volume\":\"48 \",\"pages\":\"Article 100597\"},\"PeriodicalIF\":4.1000,\"publicationDate\":\"2023-03-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"1\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"International Journal of Accounting Information Systems\",\"FirstCategoryId\":\"91\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S1467089522000495\",\"RegionNum\":3,\"RegionCategory\":\"管理学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"BUSINESS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Journal of Accounting Information Systems","FirstCategoryId":"91","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1467089522000495","RegionNum":3,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"BUSINESS","Score":null,"Total":0}
Feasibility analysis of machine learning for performance-related attributional statements
We investigate the feasibility of machine learning methods for attributional content and framing analysis in corporate reporting. We test the performance of five widely-used supervised machine learning classifiers (naïve Bayes, logistic regression, support vector machines, random forests, decision trees) in a top-down three-level hierarchical setting to (1) identify performance-related statements; (2) detect attributions in these; and (3) classify the content of the attributional statements. The training set comprises manually coded statements from a corpus of management commentary reports of listed companies. The attributions include both intra- and inter-sentential attributional statements. The results show that for both intra- and inter-sentential attributions, F1-scores of our most accurate classifier (i.e., support vector machines) vary in the range of 76% up to 94%, depending on the identification, detection and classification levels and the content characteristics of attributions. Additionally, we assess the hierarchical performance of classifiers, providing insights into a more holistic classification process for attributional statements. Overall, our results show how machine learning methods may facilitate narrative disclosure analysis by providing a more efficient way to detect and classify performance-related attributional statements. Our findings contribute to the accounting and management literature by providing a basis for implementing machine learning methodologies for research investigating attributional behavior and related impression management.
期刊介绍:
The International Journal of Accounting Information Systems will publish thoughtful, well developed articles that examine the rapidly evolving relationship between accounting and information technology. Articles may range from empirical to analytical, from practice-based to the development of new techniques, but must be related to problems facing the integration of accounting and information technology. The journal will address (but will not limit itself to) the following specific issues: control and auditability of information systems; management of information technology; artificial intelligence research in accounting; development issues in accounting and information systems; human factors issues related to information technology; development of theories related to information technology; methodological issues in information technology research; information systems validation; human–computer interaction research in accounting information systems. The journal welcomes and encourages articles from both practitioners and academicians.