Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Inde → 2006 → Prediction of weather parameters using harmonic analysis and artificial neural networ

Kulshrestha,Manjusha (2006)

Prediction of weather parameters using harmonic analysis and artificial neural networ

Kulshrestha,Manjusha

Titre : Prediction of weather parameters using harmonic analysis and artificial neural network

Auteur : Kulshrestha,Manjusha

Université de soutenance : Maharaja Sayajirao University of Baroda

Grade : Doctor of Philosophy (PhD) in Applied Mathematics 2006

Résumé
This thesis, addresses the problem of prediction of weather Parameters like, i) Prediction of annual rainfall. ii) Prediction of return period of occurred highest one-day maximum rainfall. iii) Prediction of weekly rainfall probabilities. iv) Prediction of Hourly air temperatures for a day. v) Prediction of soil temperatures at 5- 20 cm. Depths, for Anand station by invoking various methods like Artificial Neural Network, Double Fourier Series , Return Period Analysis, Gamma Distribution Model, William and Logan Model and Harmonic Analysis. The above methods are first time used for prediction of weather parameters in the Gujarat Double Fourier Series approach is first time applied to the problems of prediction in Meteorology. Results of these methods are compared with Artificial Neural Network technique by finding Root Mean Squared Error (RMSE). It turned out that ANN is the best-fit algorithm for prediction of Weather Parameters. We have also obtained a convergence result for multi input, multi output McCulloch Pitts type Neural Network by using fixed-point theorem.

Présentation et version intégrale (Shodhganga)

Page publiée le 13 avril 2021