Dataset for ESG Text Classification Using Naïve Bayes and Bag of Words to Support the Green Economy

Description

This dataset contains annotated textual data used to develop and evaluate a machine learning model for classifying Environmental, Social, and Governance (ESG) information to support green economy assessment. The data were collected from corporate annual reports, sustainability reports, and publicly available company news, then processed through text extraction, cleaning, tokenization, stop-word removal, and transformation into a Bag of Words (BoW) representation. Each document is manually labeled into ESG-related classes to enable supervised learning using the Multinomial Naïve Bayes algorithm. The dataset supports reproducibility of the model development and evaluation process reported in the associated study, including accuracy, precision, recall, and F1-score measurement. It is intended for reuse in research related to ESG text classification, sustainability reporting analysis, machine learning applications in corporate governance studies, and further methodological development using advanced natural language processing techniques.

Institutions

• Binus University

  Jakarta, Jakarta

Categories

Licence