Glossary

Supervised Learning

Suppose we have paired labelled dataset \(\{(\mathbf{x}_1, \mathbf{y}_1), (\mathbf{x}_2, \mathbf{y}_2), ..., (\mathbf{x}_n, \mathbf{y}_n)\}\), where the observation \(mathbf{x}_i\) contains \(m\) elements and the target \(\mathbf{y}_i\) contains \(c\) elements. For context, \(m\) could represent the number of unique genome sequences in our dataset, and \(c\) could be the number of unique class labels. Supervised learning is interest in learning a function that approximates the mapping from our observation space to our target space.

Observation

An observation is the data we have access to which would be an input to a machine learning model. For example, it could be data regarding genome sequences. It is normally a row in a dataset.

Target

A target is the meta data relating to an observation. For example, it could be STX type, phage type, or region of origin for genome sequence datasets like ours.

Features

The blanket term for what the context of the values represent in our dataset. For our genome sequence dataset, it is the counts of certain \(k\)-mers and it is the context of the columns of a dataset.

Metrics

We evaluate how well a model explains a dataset by examining a statistical metric. Common metrics are, Recall, Precision, and F1. See Wikipedia for some of these definitions.

k-mer

Sequencing fragments of length k, generated from splitting up assemblies or reads into smaller fragments.

unitig

High-confidence contigs of overlapping k-mers, generated from a search of a De Brujin graph for groups of overlapping fragments that together make a sequence that does not overlap with conflicting sequences.