A combined method of Adaline and KNN for rainfall forcasting

Rainfall forecasts can greatly prevent rain damageand improve agricultural and horticultural activities, tourismdevelopment, and transportation. Machine learning and statisticalmethods are used in precipitation prediction researchthat Neural networks have been widely used to. In this paper,we present a combined method involving the KNN (K nearestneighbor) classification and the Adaline neural network to predictthe amount of monthly rainfall. In this way, using the KNNalgorithm, we reduce the effect of values with a greater distancefrom the target value in the training phase and improve theresults obtained from the Adaline neural network. We use thecollected data on the amount of rainfall during ۲۰۰۶-۲۰۱۶ in”Kota Denpasar” and predict the amount of precipitation in ۲۰۱۶and compare its results with the results of the simple Adelineneural network method. Comparing the results, we concludethat the proposed combined method gives better performancecompared to the simple Adeline method.

Keywords:

rainfall forcast , Artificial neural network , Adaline , K nearest neighbor