Deep learning model for temperature prediction: an empirical study

Planning the daily routines of human life depends heavily on the weather. Knowing the weather ahead of time substantially aids in better planning for aviation, agriculture, tourism, and other operations, which avoids financial loss and casualties. People in today's technological era heavily rely on weather forecasts. Researchers and computer scientists are paying special attention to machine learning (ML) techniques in an effort to develop and adopt a different approach to the conventional way of weather prediction. Predicting the weather is difficult owing to the non-linear link between input data and output conditions. Multivariate polynomial regression (MPR) and Deep neural networks (DNN)-based models are an alternative of costly and complex traditional systems. To predict maximum temperature, deep neural network-based weather forecasting models are quite simple and can be designed with less effort and cost in comparison to a traditional forecasting system. This research work objective is to investigate and predict New Delhi’s temperature in 6-h intervals for the upcoming year using the time series dataset, using input features which include date and time, temperature, atmospheric pressure, humidity, dew point, and conditions like fog, heavy fog, drizzle, etc. In this study, ML models (MPR and DNN) are designed and implemented for temperature prediction. To evaluate the efficiency of the predictions, a comparison of the predicted temperature and the actual recorded temperature is done, and the performance and accuracy of the models are examined. The DNN model (DNNM-3) outperformed the other models with an accuracy rate of 96.4%.