BIST Sürdürülebilirlik Endeksinde Yer Alan Firmaların ESG ve Çevresel Skorlarının Makine Öğrenmesi Yöntemleriyle Tahmini
Prediction of ESG and Environmental Scores of Firms Listed in the BIST Sustainability Index Using Machine Learning Methods

Summary/Abstract: Environmental, Social, and Governance (ESG) scores serve as an important assessment tool for evaluating companies’ environmental awareness, social responsibility, and adherence to corporate governance principles. Since data related to ESG indicators are mostly published in unstructured text format, conducting systematic analyses becomes challenging from both technical and methodological perspectives. In this context, the use of Natural Language Processing (NLP) techniques in combination with text mining offers significant advantages in transforming unstructured data into meaningful features. In this study, a machine learning approach was developed to predict ESG and environmental scores with the aim of evaluating companies’ sustainability performance. The model utilized ESG and environmental scores obtained from the Refinitiv database, along with the sustainability and integrated activity reports of 95 companies included in the BIST Sustainability Index for the 2016–2023 period. A total of six different machine learning algorithms were tested on the text-based data, and it was determined that the Logistic Regression, Support Vector Machines, Gradient Boost and Light GBM algorithms achieved the highest performance in terms of evaluation metrics such as accuracy and F1 score. The applied models were found to be consistent with similar studies in the literature, and the results demonstrated the applicability of artificial intelligence-based methods in predicting ESG scores.

• Details
• Contents