Back propagation neural network : An interactive tool for effective rainfall prediction
Natural water source is one of the influencing factors for crop cultivation and
production in agricultural sector. At the same time most of the agro industrial development
interdepends the natural resources and its impact for its progress. Hence, there exists an
ever growing demand for effective rainfall prediction system. The ongoing demand for
precise forecast methods has led to thedevelopment of computerized rainfall
forecastsscenarios. This investigation makes an attempt to achieve precise rainfall
predictions using Neural Network (NN) approach. This researchintroduces a novel
exhaustive search based Maximum Frequency Weighted Feature Selection (MFWFS) using
approach to identify the significant weather parameterfor prediction.The effects of feature
selectionin model performance were also investigated. This was done by examining the
performance of the neural networksprediction model using both complete and reduced
parameters. A meticulous comparison of the overall performance indicated that the back
propagation algorithm approach based neural network model outperformed better than
existing methods. The proposed NN architecture achieved 0.0499 error rate and 95.01%
prediction accuracy using effective weather parameters. This investigation introduces
acompact interactive graphical user interface (GUI) based tool developed using C# and net
platform to enable users to conduct meteorological assessment on their own ease. This
tool enables users to train and test the NN model for various input options and to visualize
the resultsby processing the rainfall forecast scenarios.
Spanish
Chinese
Russian
German
French
Japanese
Portuguese
Hindi