Towards Vision 2047: Comparative Forecasting of India’s GDP and Top Five States GSDP Using Advanced Neural Networks

Abstract

The Gross Domestic Product (GDP) or Gross State Domestic Product (GSDP) is the key indicator used to measure the economy of a nation, as well as major states. India and a few other states have official economic vision documents such as Viksit Bharat 2047, Maharashtra Vision 2047, Viksit Uttar Pradesh @ 2047, etc. In this study, advanced neural networks have been used to project the GDP/GSDP of the nation, as well as for the top five states, up to 2047. These projections are compared against the official economic vision. The analysis reveals the feasibility and timeline towards trillion-dollar and multi-trillion-dollar economies. Models like Feedforward (FFNN), Recurrent (RNN), Long Short-Term Memory (LSTM), and Bi-LSTM have been compared with an accuracy test. The results show that the Bi-LSTM model performs better compared to all other models. The model projects India’s GDP to rise from USD 3.64 trillion in 2024–25 to nearly USD 22 trillion by 2047–48. In States Maharashtra will reach USD 2.42 trillion by 2047-48, and other states, Tamil Nadu, Uttar Pradesh, Karnataka, and Gujarat, will reach around USD 1.5 to 1.7 trillion by 2047-48. This work provides data-driven insights for policymakers, highlighting gaps between India\'s and states\' visions of GDP/GSDP versus data-driven forecasts.

Introduction

Gross Domestic Product (GDP) and Gross State Domestic Product (GSDP) are key measures of national and state economic performance. India’s long-term vision, Viksit Bharat 2047, aims for a $30 trillion economy, with states like Maharashtra, Uttar Pradesh, Tamil Nadu, Karnataka, and Gujarat setting their own targets ranging from $1 trillion to $6 trillion by 2047.

This study uses historical GDP/GSDP data (1980–2023) and applies advanced machine learning models—Feed-Forward Neural Networks (FFNN), Recurrent Neural Networks (RNN), LSTM, and Bidirectional LSTM (Bi-LSTM)—to forecast long-term economic growth up to 2047–48. Bi-LSTM outperforms other models, achieving the lowest errors (RMSE, MSE, MAE, MAPE), showing superior accuracy for long-term projections.

Results indicate Maharashtra leads in GSDP (Rs. 40.56 trillion, 2023–24), followed by Tamil Nadu, Uttar Pradesh, Karnataka, and Gujarat, collectively contributing nearly 50% of India’s economy. Forecasts suggest that Bi-LSTM can effectively guide policymakers in assessing the feasibility of state and national economic goals, highlighting the value of deep learning for complex, nonlinear economic time series.

Conclusion

For comparing the projections with the targets, the projected data from Rupees trillion is converted into USD trillion. The Bi-LSTM model projects India’s GDP to grow significantly from USD 3.64 trillion in 2024–25 to USD 21.99 trillion by 2047–48. At the state level, Maharashtra is expected to reach USD 1.04 trillion by 2034–35 and USD 2.42 trillion by 2047–48. Tamil Nadu is forecast to reach USD 1.05 trillion by 2040–41 and further increase to USD 1.63 trillion by 2047–48. Uttar Pradesh is projected to become a trillion-dollar economy by 2040–41, reaching USD 1.64 trillion by 2047–48. Karnataka is anticipated to reach USD 1.01 trillion by 2040-41 and grow to USD 1.52 trillion by 2047–48, while Gujarat is expected to attain USD 1.05 trillion by 2040–41 and USD 1.6 trillion by 2046–48.

References

[1] https://cdnbbsr.s3waas.gov.in/s3kv05b90e8875c131d0760d04c0d4d8f3/uploads/2025/08/2025080227.pdf [2] https://economictimes.indiatimes.com/news/economy/finance/maharashtra-aims-to-become-5-trillion-economy-by-2047-devendra-fadnavis/articleshow/121382466.cms?from=mdr [3] https://invest.up.gov.in/wp-content/uploads/2025/08/5-How_300825.pdf [4] https://tnidb.tn.gov.in/media/filer_public/93/ea/93ea15fb-371e-4fbe-8576-a946bb992943/tamil_nadu_vision_1_trillion_report-compressed.pdf [5] https://planning.karnataka.gov.in/storage/pdf-files/Latest%20News/Karnataka%202022-One%20Trillion%20GDP%20Vision-Mohandas%20Pai%20Nisha%20Holla.pdf [6] https://gad.gujarat.gov.in/Personnel/images/pdf/Viksit_Gujarat_2047.pdf [7] Zhang, G. P., Patuwo, B. E., & Hu, M. Y. (1998). Forecasting with artificial neural networks: The state of the art. International Journal of Forecasting, 14(1), 35–62. https://doi.org/10.1016/S0169-2070(97)00044-7. [8] Fischer, T., & Krauss, C. (2018). Deep learning with long short-term memory networks for financial market predictions. European Journal of Operational Research, 270(2), 654–669. https://doi.org/10.1016/j.ejor.2018.04.030 [9] Bhatnagar, P., & Ghosh, S. (2020). Application of recurrent neural networks for macroeconomic forecasting: Evidence from India. Journal of Forecasting, 39(7), 1105–1120. https://doi.org/10.1002/for.2702 [10] Schuster, M., & Paliwal, K. K. (1997). Bidirectional recurrent neural networks. IEEE Transactions on Signal Processing, 45(11), 2673–2681. https://doi.org/10.1109/78.650093. [11] Chen, Y., Xu, L., & Zhang, Y. (2020). Bidirectional LSTM for time series forecasting: An empirical evaluation. Applied Soft Computing, 96, 106540. https://doi.org/10.1016/j.asoc.2020.106540

Copyright

Copyright © 2026 Dr. P. Preethi, Dr. S. A. Jyothi Rani, Dr. V. V. Haragopal. This 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.