/NASA-website-queries-prediction

A hybrid approach was developed to predict NASA website queries using neural networks and metaheuristic optimization algorithms. The weights of the model was optimized using GWO, PSO, and ICA, harnessing the strengths of these algorithms to achieve remarkable results.

Primary LanguageMATLABMIT LicenseMIT

NASA website queries prediction

Here, we forecast the amount of queries to the NASA website. After fitting an MLP to the dataset, the neural network weights are retrieved. These weights are optimized using three distinct optimization algorithms: grey wolf optimization (GWO), particle swarm intelligence (PSO), and imperialist competitive algorithm (ICA). Using these three techniques, the mean absolute percentage error (MAPE) on the test set is seen to decrease.