Weather Forecasting using Bipolar Fuzzy Neural Networks

Nowadays, weather condition is unstable and weather prediction is vitally important in numerous areas of society, particularly agriculture and water management. Weather forecasting is the application of current technology and science to predict the state of the atmosphere. It is carried out by collecting as much data as possible on the meteorological characteristics about the current state of the atmosphere and using an understanding of atmospheric processes predict how the atmosphere will evolve. This paper deals with prediction for the chance of rain using bipolar fuzzy adaline and madaline network based on the temperature, wind pressure, precipitation and humidity. Input weights are considered as bipolar fuzzy sets. In Bipolar fuzzy adaline neural network, bipolar sigmoid activation function is used whereas in bipolar fuzzy madaline neural network, binary and bipolar sigmoid activation functions are used to determine the output value. If the output value is not equal to the predicted value, error is detected and the weights and bias values are updated so as to reduce the error value, until the predicted value is attained.