Uncertainty-Aware Functional Analysis for Electricity Consumption Prediction Using Multi-Task Optimization Learning Model

Energy operators require the best forecasting tools for efficient power system planning and operation. Machine learning approaches are more popular for the modelling and optimization of real-world issues with a great degree of accuracy and certainty. Variation of neural network method for predicting electric energy usage is presented and the predictive modelling is applied to a time series dataset that includes electric consumption data captured over a certain time period. The work presents a methodology that employs a Multi-task Optimization Learning Model to forecast future electricity consumption while also taking into account prior electricity usage, without compromising the data's integrity. The research focuses on investigating the performance of different Neural Network models while improving the performance through optimization technique. The proposed forecasting model outperforms the other existing model with R2 of 0.96.