Glossary of terms

Domain expertise: Specialised knowledge and/or skills in a particular field.

Epoch: One pass of the training data through the algorithm, i.e. one iteration of learning.

Learning rate: How quickly the model learns the data.

Object-oriented: A programming methodology in which the system comprises one or more modular objects that are manipulated to perform a task.

Quantization error: A SOM performance metric that measures how close each data point is to its closest neuron in the SOM.

Similarity map: A color map representing the similarity between pixels in a hyperpsectral image using color.

Topographic error: A SOM performance metric that measures how well the topology of the SOM matches the topology of the data.