Effective LSTM with K-means Clustering Algorithm for Electricity Load Prediction

Short-term electricity load has the characteristics of timing and non-linearity, which is affected by many factors such as temperature, humidity. This paper proposes a hybrid prediction algorithm based on K-means clustering and long short-term memory (LSTM). K-means, which clusters the highest temperature, the lowest temperature, humidity and other characteristics of the electricity load, divides the data set into K classes. LSTM is utilized to solve nonlinear regressive and time series problem. Simulation results based on real load data show its priority to LSTM without clustering.