Short-term forecasting of a univariate time series load using LSTM based RNN model

Due to the recent developments of the society and economy, load forecasting is becoming a crucial asset in the field of power system dispatch and demand response. We have analyzed the hourly load demand of a univariate time series. First, a single data point is predicted implementing a single-layered LSTM based RNN model. Later on, forty data points are predicted using this model using different optimizers. The experimental results prove the effectiveness of the developed prediction algorithm.