Optimized Hybrid Neural Network for Wind Speed Forecasting

Abstract

Though wind power capacity all over the world is increasing rapidly, the availability of wind power generation mostly reclines on wind speed, which is a random variable with stochastic nature. Therefore, robust technique with powerful feature extraction capability is required to predict wind speed accurately. In this paper, we have recommended a hybrid model using convolutional neural network (CNN) and long-short term memory (LSTM). where CNN is used for extracting fuzzy input features and LSTM to catch the sequence to predict wind speed accurately. As deep learning models are associated with multiple hyper-parameters with great impact, Bayesian optimization algorithm is used for hyper-parameter tuning. Additionally, the performance of some established machine learning models are added on the same data-set. It is observed that, the proposed Bayesian optimized CNN-LSTM hybrid model surpasses the other four established models like SVM, ANN, CNN and LSTM in terms of different performance evaluation metrics like mean absolute error, root mean error and root mean square error.