Machine Learning keywords

Definition of machine learning

Extracting information from data to make predictions or decisions using algorithms that build models (based on training data). These models make predictions for new & unseen data.

What are the types of machine learning?

1. Supervised

2. Unsupervised

3. Semi-supervised

What is supervised machine learning?

Uses input variables (x) and output variable (y) and build a model based on labelled data to make predictions for new data

What is unsupervised machine learning?

Only uses input variable (x), there are no output labels. The algorithm identifies patterns or groupings within the data .

What is semi-supervised machine learning?

A combination of supervised & unsupervised when labelling is expensive or impractical. There is a large amount of input data (x) but only some output data is labelled (y).

What is feature selection?

Removing non-informative or redundant predictors from the model.

What is a classification problem?

A type of supervised learning machine where the goal is to predict the class or category label of new observations based on training data with known labels. E.g. predicting if someone is likely to be promoted or not based on their education level.

What is a regression problem?

A type of supervised learning where the goal is to predict a continuous numerical outcome based on input features. E.g estimating income based on an employee’s education

What is overfitting?

Overfitting is when the model learns not only the relationships among the data, but also the noise in the data. So, it will perform well within training data set, but not on test data.