00000ctm a22000004a 4500 UP-99796217611212678 Buklod 20230215105418.0 m |o d | ta 140425s xx d r |||| u| (iLib)UPD-00215704494 DENGII eng DMLUC LG 993.5 2011 E64 V35 Valdez, Nicanor Marco P. Optimal feature extraction for Philippine ethnic gong music classification Nicanor Marco P. Valdez. 2011. xii, 96 leaves ill. + 3 DVD-ROMs (4 3/4 in.) Undergraduate student project (B.S. Electronics and Communications Engineering)--University of the Philippines, Diliman. Before a classification program can differentiate between the music of the different Philippine Ethnic Groups, the musical features that provide the basis for differentiation must be obtained. This project used forward Sequential Feature Selection (SFS) to find the optimal features for Philippine Ethnic Gong Music Classification. The accuracy ratings of two Classification systems (Support Vector Machine (SVM) and Neural Network) were used as the SFS criterion. Features used reflected the musical attributes Timbre, Loudness, Rhythm, and Pitch-and-Melody. K-Means Clustering was used on the best features to evaluate the similarity between the gong music styles of the Philippine Ethnic Groups used in the study. The 400 audio samples used in the study came from the archive of the Jose Maceda Collection of the Center for Ethnomusicology located at the UP College of Music. The Ethnic Groups used were the following: Ibaloi, Ifugao, Palawan, Tagbanua, Maguindanao, Maranao, Tiruray, Samal, Tausug and Yakan. The highest accuracy rating, 90.83%, was garnered when using SVM, the full timbre-related feature set and 30-second segments of the audio samples. The highest accuracy rating garnered using neural networks was 82.5% when the full timbre-related feature set and 30-second segments were used. These results indicate that Timbre is the most significant musical attribute when differentiating the gong musical styles of the different Philippine Ethnic Groups. Classification Music. Musical analysis Data processing. Folk music Philippines. Gong music. Thesis FI UP UPD DENG-II LG 993.5 2011 E64 V35 UPD DMUE Thesis