LSSA-BP-based cost forecasting for onshore wind power
Feng Ren, Liu Wencheng
Abstract
An LSSA-BP neural network prediction model was established for more accurate onshore wind power cost prediction. Optimise the weights and thresholds of the BP neural network using the sparrow search algorithm. Comparison of the traditional BP model, GA-BP model and LSSA-BP model to verify the superiority of the LSSA-optimised BP model. Moreover, using LSSA-BP in compared with Support Vector Regression Forecasting (SVR) and Random Forest Regression Forecasting (RFR) models. The results of model trial calculations and analysis showed that the LSSA-BP model had the highest prediction accuracy and could be used as a reference for the onshore wind power cost prediction.
Topics & Concepts
Artificial neural networkSupport vector machineRandom forestWind powerRegressionSparrowRegression analysisComputer sciencePredictive modellingPower (physics)BackpropagationMeteorologyData miningArtificial intelligenceEngineeringMachine learningStatisticsMathematicsGeographyBiologyPhysicsQuantum mechanicsEcologyElectrical engineeringAdvanced Decision-Making TechniquesAdvanced Computational Techniques and ApplicationsGeoscience and Mining Technology