Services on Demand
Journal
Article
Indicators
- Cited by SciELO
- Access statistics
Related links
- Cited by Google
- Similars in SciELO
- Similars in Google
Share
Revista EIA
Print version ISSN 1794-1237
Abstract
TOBON-GONZALEZ, Indira Juliana and CORTES-OSORIO, Jimmy Alexander. Identification of Musical Instruments of Plucked Strings of the Colombian Andean Region in Solo Using Machine Learning Techniques. Rev.EIA.Esc.Ing.Antioq [online]. 2018, vol.15, n.30, pp.177-193. ISSN 1794-1237. https://doi.org/10.24050/reia.v15i30.1245.
There are many studies on the identification of musical instruments, but none has focused on plucked string instrument of the Colombian Andean region such as the tiple, tiple requinto, guitar and bandola. Therefore, we propose to identify these instruments using machine-learning techniques such as: Discriminant Analysis, Decision Tree, k-Nearest Neighbors (kNN), Support Vector machines (SVM), Artificial Neural Network (ANNs) and three methods of data reduction: Feature Selection; Principal Components Analysis (PCA) with 10,100 and 1000 principal components, and extracting the first five partial frequencies along with their normalized amplitudes. We carried out this study using a database of 1000 digital monophonic audio recordings, built of the recordings of the first position of the notes played in solo for each string instrument in WAV format. Regarding the validity method, the Cross Validation Method was used with a k equal to five to perform the confusion matrices and the ROC Curves (Receiver Operating Characteristic). We reached the best results with ANNs that had an accuracy of 99.8%, besides the ROC curves showed an area under the curve very close to one for the guitar.
Keywords : Machine Learning; Fourier; Identification; Musical Instruments; Confusion Matrix; Colombian Andean region.