AI-Driven Portfolio Optimization for ESG Investment
DOI:
https://doi.org/10.63282/3050-922X.IJERET-V3I4P114Keywords:
Environmental, Social And Governance (ESG), Artificial Intelligence (AI), Portfolio Optimization, Natural Language Processing (NLP), Large Language Models (Llms), Sustainable Investing, Financial Text Mining, Earnings Call Analysis, Regulatory Filings (10-K), ESG Signal Extraction, Multi-Objective OptimizationAbstract
Environmental Social and Governance (ESG) considerations are fast becoming part and parcel of investment strategy. Meanwhile, artificial intelligence (AI) techniques, namely, natural language processing (NLP) and large language models (LLMs) are opening up the field to in-depth analysis of unstructured corporate disclosures and financial texts. This paper gives a model of an AI-based portfolio optimization taking into account the ESG factors becoming a direct part of the decision-making process. The methodology combines systematic historical market data together with unstructured textual evidence taken out of earnings calls, regulatory filings, and sustainability reports. The pipeline includes three major components: 1. ESG signal extraction by means of domain-adapted transformers; 2. Structured fact cogeneration and summarization with LLMs; 3. Multi-objective portfolio optimization that incorporates balancing returns, risk, and ESG alignment. The model is set with two conference contexts in view. In the track of ICDM, the task is to mine ESG-related information on earnings calls and 10-K reports using mining based on NLP. In DSAA track, the interest is in the use of LLMs to process financial documents to enable ESG assessment that can be audited. The framework is grounded conceptually in Adewale (2024), which argues for integrating ESG criteria in AI-based portfolio management, and in Rane, Choudhary, and Rane (2024), which surveys AI approaches that strengthen ESG in sustainable business practices. The proposed solution shows the promise of AI-based source of investments in giving sustainability-focused investing an edge on appropriate textual evidence to be converted into explainable and actionable portfolio allocation signals.
References
[1] Adewale, T. (2024). Integration of ESG (Environmental, Social, Governance) Criteria into AI-Based Portfolio Management.
[2] Castelberg, D., & Flury, F. (2024). Llm assisted development (Doctoral dissertation, OST Ostschweizer Fachhochschule).
[3] Chen, C. C., Huang, H. H., & Chen, H. H. (2020). NLP in FinTech applications: past, present and future. arXiv preprint arXiv:2005.01320.
[4] Joshi, R., Bubna, Y., Sahana, M., & Shruthiba, A. (2024, November). An Approach to Intelligent Information Extraction and Utilization from Diverse Documents. In 2024 8th International Conference on Computational System and Information Technology for Sustainable Solutions (CSITSS) (pp. 1-5). IEEE.
[5] Kannan, N., & Seki, Y. (2023). Textual evidence extraction for ESG scores. In Proceedings of the Fifth Workshop on Financial Technology and Natural Language Processing and the Second Multimodal AI For Financial Forecasting (pp. 45-54).
[6] Konieczny, K. (2024). Wisdom from Words: Evidence from Earnings Conference Calls and Analysts’ Research Reports using State-of-the-Art Natural Language Processing Methodologies (Doctoral dissertation, Technische Universität München).
[7] Nair, V. (2024). AI-powered investment strategies: Enhancing portfolio management through machine learning. Journal Of Recent Trends in Computer Science and Engineering (JRTCSE), 12(1), 1-5.
[8] Paul, C. (2023). The Role of AI in Creating Sustainable and ESG-Focused Hedge Fund Portfolios.
[9] Perazzoli, S., Joshi, A., Ajayan, S., & de Santana Neto, J. P. (2022). Evaluating environmental, social, and governance (esg) from a systemic perspective: An analysis supported by natural language processing. Social, and Governance (ESG) from a Systemic Perspective: An Analysis Supported by Natural Language Processing (October 11, 2022).
[10] Raman, N., Bang, G., & Nourbakhsh, A. (2020). Mapping ESG trends by distant supervision of neural language models. Machine Learning and Knowledge Extraction, 2(4), 453-468.
[11] Rane, N., Choudhary, S., & Rane, J. (2024). Artificial intelligence driven approaches to strengthening Environmental, Social, and Governance (ESG) criteria in sustainable business practices: a review. Social, and Governance (ESG) criteria in sustainable business practices: a review (May 27, 2024).
[12] Roger, J. D. D. (2024). AI-driven financial modeling techniques: transforming investment strategies. The Journal of Applied Business and Economics, 26(4), 63-74.
[13] Ruberg, N. (2021). Bert goes sustainable: an NLP approach to ESG financing.
[14] Sun, Z., Satapathy, R., Guo, D., Li, B., Liu, X., Zhang, Y., ... & Goh, R. S. M. (2024, December). Information extraction: Unstructured to structured for esg reports. In 2024 IEEE International Conference on Data Mining Workshops (ICDMW) (pp. 487-495). IEEE.
[15] Talapaneni, A. (2024). Enhancing ESG Reporting and SDG Mapping Using Natural Language Processing.
[16] Taraj, I. (2024). Financial and Environmental Narratives in Earnings Calls: Investor Processing Costs and Market Reactions.
[17] Wang, Y. (2024). NLP applications—clinical documents. In Natural Language Processing in Biomedicine: A Practical Guide (pp. 325-349). Cham: Springer International Publishing.
[18] Zhang, A. Y., & Zhang, J. H. (2024). Renovation in environmental, social and governance (ESG) research: the application of machine learning. Asian Review of Accounting, 32(4), 554-572.
