A Comparison Of Support Vector Machine And Fuzzy Type-2 Techniques In Weather Forecasting

Liew , Chew Hong (2011) A Comparison Of Support Vector Machine And Fuzzy Type-2 Techniques In Weather Forecasting. Project Report. UTeM, Melaka, Malaysia. (Submitted)

	PDF (24 Pages) A_Comparison_Of_Support_Vector_Machine_And_Fuzzy_Type-2_Techniques_In_Weather_Forecasting-24_pages.pdf - Submitted Version Download (3MB)
	PDF (Full Text) A_Comparison_Of_Support_Vector_Machine_And_Fuzzy_Type-2_Techniques_In_Weather_Forecasting-Full_Text.pdf - Submitted Version Restricted to Registered users only Download (15MB)

Abstract

Weather forecasting is an important application by providing accurate weather prediction to saves lives, money and time in both local and global area. Type-2 Fuzzy Logic is one of the practiced automated techniques used to forecast weather. However, the research shows that Type-2 Fuzzy Logic does not obtain good performance when number of training data is small because of inadequate of rules to build a prediction system. Hence, this work had proposed Support Vector Machines (SVM) as the automated techniques for weather forecasting which can complement the weakness of Type-2 Fuzzy Logic and compared the performance of both techniques. The dataset which were recorded every half hour interval originally was subsample into weather dataset I (3 hourly sample), weather dataset 2(2hourly sample) and weather dataset 3(daily average sample). The aim of the experiments was to investigate the performance of both techniques on the datasets with different size. The experiment environment is developed with C++ programming language in Matlab.The results of both techniques are compared in terms of Mean Squared Error (MSE) and validated using T-test. The experimental result shows that Support Vector Machine (SVM) had performed consistently better than Type-2 Fuzzy Logic due to its characteristic which is highly scalability on data with different dimensionality. Researchers are suggested to hybrid Type-2 Fuzzy Logic and Support Vector Machines (SVM) on the future work since Type-2 Fuzzy Logic had advanced in handling uncertainties of weather data.

Actions (login required)

View Item