A FUSION MODEL FOR STOCK MARKET PREDICTION USING PROPHET AND LONG SHORT-TERM MEMORY NEURAL NETWORKS
Predicting the stock market can be difficult because of its inherent volatility and complexity. Machine learning approaches have demonstrated potential in identifying patterns and trends in financial data, enabling precise prediction-making in recent times. In this work, we combine the advantages of...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
University of Kragujevac
2025-03-01
|
| Series: | Proceedings on Engineering Sciences |
| Subjects: | |
| Online Access: | https://pesjournal.net/journal/v7-n1/64.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Predicting the stock market can be difficult because of its inherent volatility and complexity. Machine learning approaches have demonstrated potential in identifying patterns and trends in financial data, enabling precise prediction-making in recent times. In this work, we combine the advantages of Prophet regression and Long Short-Term Memory (LSTM) neural networks to propose a fusion model for stock market prediction. Because the LSTM model is so good at capturing temporal relationships in sequential data, it is a great choice for studying historical trends in stock prices. Conversely, Facebook's Prophet is a powerful time-series forecasting tool that makes accurate predictions by taking into account patterns, seasonality, and holidays. Our fusion strategy takes advantage of the complimentary capabilities of both techniques by merging Prophet and LSTM models. The Prophet component takes seasonal trends and outside influences into consideration to further improve forecasts, while the LSTM component analyzes past stock market data to identify intricate patterns. We validate our fusion model's efficacy through tests on real-world stock market datasets. We compare our forecasts' accuracy to that of individual LSTM and Prophet models as well as conventional forecasting techniques. Our findings show that the fusion model performs better than stand-alone methods, resulting in increased reliability and prediction accuracy. Predicting the stock market can be difficult because of its inherent volatility and complexity. Machine learning approaches have demonstrated potential in identifying patterns and trends in financial data, enabling precise prediction-making in recent times. In this work, we combine the advantages of Prophet regression and Long Short-Term Memory (LSTM) neural networks to propose a fusion model for stock market prediction. Because the LSTM model is so good at capturing temporal relationships in sequential data, it is a great choice for studying historical trends in stock prices. Conversely, Facebook's Prophet is a powerful time-series forecasting tool that makes accurate predictions by taking into account patterns, seasonality, and holidays. Our fusion strategy takes advantage of the complimentary capabilities of both techniques by merging Prophet and LSTM models. The Prophet component takes seasonal trends and outside influences into consideration to further improve forecasts, while the LSTM component analyzes past stock market data to identify intricate patterns. We validate our fusion model's efficacy through tests on real-world stock market datasets. We compare our forecasts' accuracy to that of individual LSTM and Prophet models as well as conventional forecasting techniques. Our findings show that the fusion model performs better than stand-alone methods, resulting in increased reliability and prediction accuracy. |
|---|---|
| ISSN: | 2620-2832 2683-4111 |