Language Modeling for the Future of Finance: A Quantitative Survey into Metrics, Tasks, and Data Opportunities
This paper systematically surveys NLP applications in finance covering models, datasets, tasks, evaluation metrics, crisis data, and research trends.
What it examines
This survey systematically reviews 221 NLP papers applied to finance by categorizing tasks, datasets, methods, and evaluation metrics. It examines the shift from custom to general language models and explores trends in sentiment analysis, forecasting, and explainability, using both qualitative and quantitative analyses to map the field.
What it concludes
The study finds that general-purpose language models dominate, though challenges like crisis data and domain-specific metrics persist. Applications include enhanced market forecasting, risk management, and sentiment analysis. Future research should diversify data sources and methods to build more robust and practical financial decision-making tools.
Evidence objects
The survey reveals adoption of general-purpose NLP models like BERT and GPT-4 in finance since 2017, as researchers increasingly favor pretrained frameworks over architectures, integrating broad advances into financial tasks.
key_findings bullet 1 · key_findings · validation V0
A systematic evaluation of 221 finance-related papers examines 11 quantitative and qualitative dimensions, uncovering gaps in explainability, privacy, crisis forecasting, and offering financial task definitions to guide future research directions.
key_findings bullet 2 · key_findings · validation V0
The study observes open-source practices, challenges in code maintenance, limited non-English data and alternative method usage, and suggests replacing machine learning metrics with financial-specific measures like $$Sharpe\\ Ratio$$ for performance validation.
key_findings bullet 3 · key_findings · validation V0
Combining originality and robust analysis, the paper presents a structured quantitative survey of NLP applications in finance by evaluating trends and gaps in existing methodologies. Its systematic exploration of evaluation metrics, tasks, and datasets offers moderately compelling insights, though it does not deliver revolutionary breakthroughs, engaging readers with consolidated perspectives.
key_findings bullet 4 · key_findings · validation V0
Raw abstract and provenance
Abstract: Recent advances in language modeling have led to growing interest in applying Natural Language Processing (NLP) techniques to financial problems, enabling new approaches to analysis and decision-making. To systematically examine this trend, we review 374 NLP research papers published between 2017 and 2024 across 38 conferences and workshops, with a focused analysis of 221 papers that directly addr… ▽ More Recent advances in language modeling have led to growing interest in applying Natural Language Processing (NLP) techniques to financial problems, enabling new approaches to analysis and decision-making. To systematically examine this trend, we review 374 NLP research papers published between 2017 and 2024 across 38 conferences and workshops, with a focused analysis of 221 papers that directly address finance-related tasks. We evaluate these papers across 11 qualitative and quantitative dimensions, identifying key trends such as the increasing use of general-purpose language models, steady progress in sentiment analysis and information extraction, and emerging efforts around explainability and privacy-preserving methods. We also discuss the use of evaluation metrics, highlighting the importance of domain-specific ones to complement standard machine learning metrics. Our findings emphasize the need for more accessible, adaptive datasets and highlight the significance of incorporating financial crisis periods to strengthen model robustness under real-world conditions. This survey provides a structured overview of NLP research applied to finance and offers practical insights for researchers and practitioners working at this intersection. △ Less
Source row: 1154 · abstract type: unknown