Classification and Forecasting of Weather using ANN, k-NN and Na?ve Bayes Algorithms

Nishchala C. Barde, Mrunalinee Patole

Abstract: Weather forecasting is a way to predict future weather. It is widely researched area due to the fact that human life on earth is affected by the global climate. In this paper, we have proposed a comparative study between various techniques for prediction. This work focuses on developing an optimised system model which predicts future weather. The frequency of natuaral hazards occuring due to unpredictable weather conditions have been seen to be increasing causing damage to human life. There are some models that predict weather during real time, or monthly or annual period. This paper presents a system that carries out weather prediction using previous or historical weather data having attributes (Date, Temperature, Humidity, Wind Chill (WC) and StnPressure (SP). Various data mining techniques are used for this purpose of weatherforecasting such as Multi-layer Perceptron, K-Nearest Neighbor and naive Bayes. Comparison based on evaluation parameters identify which model has more accurately performed the predictions.

Keywords: Machine Learning, Artificial Neural Network, Nave Bayes, K Nearest Neighbors

How to Cite?: Nishchala C. Barde, Mrunalinee Patole, "Classification and Forecasting of Weather using ANN, k-NN and Na?ve Bayes Algorithms", Volume 5 Issue 2, February 2016, International Journal of Science and Research (IJSR), Pages: 1740-1742, https://www.ijsr.net/getabstract.php?paperid=NOV161672, DOI: https://dx.doi.org/10.21275/NOV161672