Project for the IAS course
(part A)
➢ Download the ESC-50 dataset (https://github.com/karolpiczak/ESC-50)
and select 3 classes of your preference (if that link does not work
download it from https://unimi2013-
my.sharepoint.com/:u:/g/personal/stavros_ntalampiras_unimi_it/EYuvdl0
SXxhBki5He8D502oBqHoS9k1GYCo-g1Wd9i5TVw?e=Cpnccm)
➢ Divide the data into 70% for training and 30% for testing.
➢ Extract all the audio features (time and frequency domain) that we have
learnt.
➢ Visualize the feature space in 3D using PCA and report the number of
coefficients which offer at least 80% of variance.
➢ Train the kNN classifier on the train set to classify the test set. Use three
groups of features: a) time, b) frequency and c) altogether.
Classification of generalized audio