Weather prediction plays a vital role in safeguarding life and optimizing resource planning. However, the inherent chaotic nature of atmospheric systems makes precise forecasting a persistent challenge. Traditional numerical and statistical models often lack adaptability and accuracy, especially in rapidly changing weather conditions. These models may not fully leverage the potential of data-driven adaptive intelligence for real-time prediction. This study proposes a novel weather prediction model based on Swarm Intelligence (SI), specifically utilizing a hybrid Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO) algorithm. The hybrid SI model is designed to fine-tune predictive parameters dynamically while adapting to spatio-temporal variations in meteorological data. The framework incorporates multi-source weather data (temperature, humidity, pressure, and wind speed) and applies optimized machine learning regression models, whose hyperparameters are tuned through the SI-based approach. The proposed SI model was tested against benchmark datasets using MATLAB simulations. It showed improved prediction accuracy and adaptability compared to existing methods, including ARIMA, Support Vector Regression (SVR), LSTM, and standalone PSO-tuned models. The hybrid SI framework achieved a notable increase in accuracy (6–12%) and reduced prediction error across different climate zones, demonstrating its effectiveness in dynamic conditions.
Rajesh K. Agrawal, Swati B. Baste, Yogita S. Rathod
SNJB’s Late Sau Kantabai Bhavarlalji Jain College of Engineering, India
Swarm Intelligence, Weather Prediction, Particle Swarm Optimization, Ant Colony Optimization, Forecast Accuracy
| January | February | March | April | May | June | July | August | September | October | November | December |
| 0 | 0 | 0 | 2 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Published By :
ICTACT
Published In :
ICTACT Journal on Soft Computing
( Volume: 16 , Issue: 1 , Pages: 3803 - 3807 )
Date of Publication :
April 2025
Page Views :
29
Full Text Views :
2
|