Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{198841,
        author = {Huzaifa Zahid Husein Shah and Mohammad Ali Ansari and Aditya Vijay Kamble and Talha Ansari and Ahlam Ansari},
        title = {Intelligent Multi-Asset Portfolio Advisory System Using Deep Learning and Natural Language Processing},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {11769-11775},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=198841},
        abstract = {Retail investors in India face growing difficulty navigating a fragmented and increasingly complex financial landscape. Existing decision-support tools typically address individual asset categories in isolation, offering little integration across multiple investment domains. This paper proposes an Intelligent Multi-Asset Portfolio Advisory System that consolidates five asset classes equities, mutual funds, precious metals, residential real estate, and digital currencies into a single analytical framework. Market data is acquired through public APIs, automated web scraping, and curated open datasets, and subsequently processed using a suite of machine learning and deep learning techniques. Structured asset data for metals and real estate is modelled via Random Forest and Prophet, while Long Short-Term Memory (LSTM) networks perform time-series forecasting for equities, digital currencies, and mutual fund NAVs. A FinBERT-based module performs contextual sentiment classification on financial news, and a large language model layer translates model outputs into plain-language advisory narratives. The system achieved R² values of 0.90 for gold, 0.88 for residential real estate, 0.78 for equities, 0.85 for digital currencies, and 0.99 for mutual funds. Experimental results indicate that integrating quantitative forecasting with sentiment-aware natural language generation substantially improves both predictive accuracy and user interpretability, advancing the accessibility of data-driven financial guidance for everyday investors.},
        keywords = {Deep Learning, Digital Currencies, Equities Forecasting, FinBERT, LSTM, Multi-Asset Framework, Mutual Funds, Natural Language Advisory, Portfolio Intelligence, Real Estate Analytics},
        month = {April},
        }