Application of Data Mining for Weather Forecasting
No Thumbnail Available
Date
2015-01-03
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
Weather forecasting has been one of the most scientifically and technologically challenging
problems around the world in the last century. Making an accurate prediction is one of the major
challenges meteorologist are facing all over the world. Since ancient times, weather prediction
has been one of the most important domains. Scientists have tried to forecast meteorological
characteristics using a number of methods, some of these methods being more accurate than
others.
Accurate and timely weather forecasting is a major challenge for the National Meteorological
Agency of Ethiopia. In this study, the researcher investigated the use of data mining techniques
in forecasting rainfall. This was carried out using J48 decision tree, Multilayer perceptron
artificial neural network, and PART rule induction algorithms and meteorological data collected
between 2000 and 2014 from National Meteorological Agency of Ethiopia. A data model for the
meteorological data was developed and this was used to train the classifier algorithms. The
performances of these algorithms were compared using standard performance metrics, and the
algorithm which gave the best results used to generate classification rules for the mean weather
variables. A predictive model was also developed for the weather prediction program and the
results compared with actual weather data for the predicted periods. The results show that given
enough case data, Data Mining techniques can be used for weather forecasting.
To get a better awareness in choosing which model produced sound prediction and higher
accuracy, 13 experiments were done with J48 algorithm and multilayer perceptron classifier, and
eight experiments were done using PART rule induction, by inputting all the records with a 10
fold cross-validation mode, and inputting different percentage (%) of the record for testing the
performance of the model. The next option used by the researcher to improve the performance of
the model were to test if a better model could be obtained by excluding one or more of the input
variables and training different models. J48 has an accuracy of 86.65%, PART has an accuracy
of 84.96 and Neural Network has an accuracy of 80.03%. Then J48 algorithm has shown better
prediction performance. In the future, the effective use of information and technology is
important for National Meteorology to stay competitive in today’s complex environment. The
challenges faced when trying to make large, diverse, and often complex dataset records are
considerable and employing other classification algorithms could yield better results.
Description
Keywords
Application of Data Mining