Financial Market Sentiment Analysis Using Artificial Intelligence Techniques

Abstract

Financial market sentiment analysis has emerged as a crucial tool for understanding investor behavior and predicting market movements. The combination of AI and NLP has greatly improved the degree and quality of sentiment-based predictions of the financial institutions. This paper presents a comprehensive survey of sentiment analysis and turns from lexicon-based approaches to deep learning models like LSTM, BERT, and FinBERT. The proposed steps are text preprocessing, feature extraction and selection, modelling and classification, and performance measurement using precision, recall, F1-score or ROC-AUC. Compared with the other methods, one sees that FinBERT has better performance of the classification of sentiments; this proves to improve the results of sentiment based trading strategies. Indeed, the study provides substantial support for the role which sentiment analysis offers through artificial intelligence in investment, credit risk, and other applications of trading algorithms. Future advancements should focus on enhancing model interpretability, scalability, and real-time sentiment tracking for improved financial forecasting.