Impact of Hybrid Intelligent Computing in Identifying Constructive Weather Parameters for Modeling Effective Rainfall Prediction
No hay miniatura disponible
Fecha
Autores
Título de la revista
ISSN de la revista
Título del volumen
Editor
Resumen
Descripción
Uncertain atmosphere is a prevalent factor affecting the existing prediction approaches. Rough set
and fuzzy set theories as proposed by Pawlak and Zadeh have become an effective tool for handling vagueness
and fuzziness in the real world scenarios. This research work describes the impact of Hybrid Intelligent
System (HIS) for strategic decision support in meteorology. In this research a novel exhaustive search based
Rough set reduct Selection using Genetic Algorithm (RSGA) is introduced to identify the significant input
feature subset. The proposed model could identify the most effective weather parameters efficiently than
other existing input techniques. In the model evaluation phase two adaptive techniques were constructed
and investigated. The proposed Artificial Neural Network based on Back Propagation learning (ANN-BP)
and Adaptive Neuro Fuzzy Inference System (ANFIS) was compared with existing Fuzzy Unordered Rule
Induction Algorithm (FURIA), Structural Learning Algorithm on Vague Environment (SLAVE) and Particle
Swarm OPtimization (PSO). The proposed rainfall prediction models outperformed when trained
with the input generated using RSGA. A meticulous comparison of the performance indicates ANN-BP model
as a suitable HIS for effective rainfall prediction. The ANN-BP achieved 97.46% accuracy with a nominal
misclassification rate of 0.0254 %.