A hybrid CNN-LSTM approach to enhancing temperature forecasting for environmental threats and risk management

Temperature is the most important element of weather, which is applicable in varied study areas such as environmental, ecological, industry, agriculture sectors, etc. This research platforms the practicality of utilising a combination of convolutional neural networks and learning paradigms to forecast weather conditions in the eastern region of India, New Delhi. The authors propose long short-term memory (LSTM) networks, convolutional neural networks (CNNs) and examine how they compare to hybrid CNN-LSTM model for temperature forecasting. Our aim is to address these issues through a representation, which jointly predicts temperature over time. Experiments on actual meteorological data used in our evaluation of the models highlight the approach's potential. We also used accuracy, mean square error (MSE), and root mean square error (RMSE) to estimate these models' outcome. Our findings demonstrate that the proposed CNN-LSTM model delivers the best outcomes because of its accuracy and small error rates.