Combining many features in a classifier
You've spent this lesson engineering many features from the audio data - some contain information about how the audio changes in time, others contain information about the spectral content that is present.
The beauty of machine learning is that it can handle all of these features at the same time. If there is different information present in each feature, it should improve the classifier's ability to distinguish the types of audio. Note that this often requires more advanced techniques such as regularization, which we'll cover in the next chapter.
For the final exercise in the chapter, we've loaded many of the features that you calculated before. Combine all of them into an array that can be fed into the classifier, and see how it does.
This exercise is part of the course
Machine Learning for Time Series Data in Python
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Loop through each spectrogram
bandwidths = []
centroids = []
for spec in spectrograms:
# Calculate the mean spectral bandwidth
this_mean_bandwidth = np.____(lr.feature.____(S=spec))
# Calculate the mean spectral centroid
this_mean_centroid = np.____(lr.feature.____(S=spec))
# Collect the values
bandwidths.append(this_mean_bandwidth)
centroids.append(this_mean_centroid)