0.4. Data modeling

Definitions

Term	Definition
features	descriptive numerical representations to describe an object.
machine learning	the process of teaching a machine something that is useful.
classification model	If the goal is to separate out into classes (e.g. male or female), then this is known as a classification problem.
label	the tag of an audio sample (e.g. male or female) to aid in supervised learning.
regression model	if the end goal is to measure some correlation with a variable and the output is more a numerical range (e.g. often between 0 and 1), then this is more of a regression problem.
unsupervised learning	f machines do not require labels (e.g. just need features), this is known as a unsupervised learning problem.
supervised learning	if machines require labels (e.g. male or female as separate feature arrays), this is known as a supervised learning problem.
cross-validation	how the performance of ML models are assessed (in terms of accuracy).
deep learning	machines that are trained using a neural network.

4.2 - Obtaining training data

Sample text

4.3 - Labeling training data

Sample text

4.4 - Classification models

Sample text

4.5 - Regression models

Sample text

4.6 - Deep learning models

Sample text

4.7 - AutoML approaches

Sample text