{"title":"哪种观点更值得信赖?基于机器学习的分析师盈利预测质量评估框架","authors":"Yingying Song , Xinxin Chen","doi":"10.1016/j.najef.2024.102318","DOIUrl":null,"url":null,"abstract":"<div><div>Analysts’ Earnings Forecast (AEF) is a crucial reference in investment decision-making and significantly impact capital market efficiency. While much research has focused on the factors influencing AEF, the variability and disparity in its quality have often been overlooked. This study presents a machine learning (ML)-based framework for assessing and forecasting AEF quality, including multi-perspective feature generation, rank aggregation-based heterogeneous ensemble feature selection, and quality forecasting. We validate this framework on a real-world dataset and use an explainable approach to identify the key features affecting AEF quality from a data-driven perspective. Our analyses reveal the unique characteristics of the China’s A-share market in terms of AEF quality forecasting and investigate the sensitivity of feature combinations from the perspectives of state ownership and industry. On the basis of our assessment, we develop an investment strategy to demonstrate economic value. Our findings offer insights for regulators and brokerage houses, helping investors mitigate the risks associated with low-quality opinions.</div></div>","PeriodicalId":47831,"journal":{"name":"North American Journal of Economics and Finance","volume":"75 ","pages":"Article 102318"},"PeriodicalIF":3.8000,"publicationDate":"2024-11-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Which opinion is more trustworthy: An analysts’ earnings forecast quality assessment framework based on machine learning\",\"authors\":\"Yingying Song , Xinxin Chen\",\"doi\":\"10.1016/j.najef.2024.102318\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><div>Analysts’ Earnings Forecast (AEF) is a crucial reference in investment decision-making and significantly impact capital market efficiency. While much research has focused on the factors influencing AEF, the variability and disparity in its quality have often been overlooked. This study presents a machine learning (ML)-based framework for assessing and forecasting AEF quality, including multi-perspective feature generation, rank aggregation-based heterogeneous ensemble feature selection, and quality forecasting. We validate this framework on a real-world dataset and use an explainable approach to identify the key features affecting AEF quality from a data-driven perspective. Our analyses reveal the unique characteristics of the China’s A-share market in terms of AEF quality forecasting and investigate the sensitivity of feature combinations from the perspectives of state ownership and industry. On the basis of our assessment, we develop an investment strategy to demonstrate economic value. Our findings offer insights for regulators and brokerage houses, helping investors mitigate the risks associated with low-quality opinions.</div></div>\",\"PeriodicalId\":47831,\"journal\":{\"name\":\"North American Journal of Economics and Finance\",\"volume\":\"75 \",\"pages\":\"Article 102318\"},\"PeriodicalIF\":3.8000,\"publicationDate\":\"2024-11-22\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"North American Journal of Economics and Finance\",\"FirstCategoryId\":\"96\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S1062940824002432\",\"RegionNum\":3,\"RegionCategory\":\"经济学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"BUSINESS, FINANCE\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"North American Journal of Economics and Finance","FirstCategoryId":"96","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1062940824002432","RegionNum":3,"RegionCategory":"经济学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"BUSINESS, FINANCE","Score":null,"Total":0}
Which opinion is more trustworthy: An analysts’ earnings forecast quality assessment framework based on machine learning
Analysts’ Earnings Forecast (AEF) is a crucial reference in investment decision-making and significantly impact capital market efficiency. While much research has focused on the factors influencing AEF, the variability and disparity in its quality have often been overlooked. This study presents a machine learning (ML)-based framework for assessing and forecasting AEF quality, including multi-perspective feature generation, rank aggregation-based heterogeneous ensemble feature selection, and quality forecasting. We validate this framework on a real-world dataset and use an explainable approach to identify the key features affecting AEF quality from a data-driven perspective. Our analyses reveal the unique characteristics of the China’s A-share market in terms of AEF quality forecasting and investigate the sensitivity of feature combinations from the perspectives of state ownership and industry. On the basis of our assessment, we develop an investment strategy to demonstrate economic value. Our findings offer insights for regulators and brokerage houses, helping investors mitigate the risks associated with low-quality opinions.
期刊介绍:
The focus of the North-American Journal of Economics and Finance is on the economics of integration of goods, services, financial markets, at both regional and global levels with the role of economic policy in that process playing an important role. Both theoretical and empirical papers are welcome. Empirical and policy-related papers that rely on data and the experiences of countries outside North America are also welcome. Papers should offer concrete lessons about the ongoing process of globalization, or policy implications about how governments, domestic or international institutions, can improve the coordination of their activities. Empirical analysis should be capable of replication. Authors of accepted papers will be encouraged to supply data and computer programs.