AI-Driven Sentiment Analysis with Natural Language Processing (NLP) for Enhanced Property Market Valuation in Malaysia

What the paper says

This study aims to enhance property market valuation in Malaysia by integrating AI-driven sentiment analysis with traditional econometric models and advanced machine learning (ML) techniques. The proposed framework combines ARIMA and GARCH models with deep learning algorithms such as Long-Short Terem Memory (LSTM) and Artificial Neural Network (ANN). Sentiment analysis leverages Natural Language Processing (NLP) through Transformer-based models, particularly Bidirectional Encoder Representation from Transformers (BERT), to extract market sentiment from property-related news, social media discussions, and corporate disclosures. The results demonstrate that this approach significantly improves valuation accuracy by capturing historical trends alongside real-time sentiment shifts. By merging sentiment analysis with property market data, the study offers a more comprehensive understanding of market dynamics, enabling more accurate predictions of price fluctuations and market movements. The research’s novelty lies in the integration of qualitative sentiment insights with quantitative property indicators, facilitating the timely detection of market changes and better insights into investor behaviour. Additionally, the model exhibits enhanced fairness and adaptability, ensuring consistent performance across Malaysia’s multilingual and diverse real estate market. The findings provide valuable implications for property investors, developers, and policymakers, as the enhanced models support better investment strategies, effective risk management, and informed decision-making in Malaysia’s evolving property landscape.