Classifying with small models · Module E of 6
Exploring data by hand
What is actually in the data before I label it?
Projection, clustering, neighbours and topic models as tools for looking, plus a labelling workflow that uses them.
Topics in this module
Topics
🔒 Unlocks at 1 of 1 topics in Module C — 0 done.- E1
Dimensionality reduction Planned
PCA, t-SNE and UMAP: look, don’t measure.
- E2
Clustering Planned
K-means versus HDBSCAN; clusters are not labels.
- E3
k-nearest neighbours Planned
Label by the neighbours’ vote, and use it to audit mislabels.
- E4
Topic modelling with NMF Planned
Readable word topics from TF-IDF; weak on short text.
- E5
Gaussian mixture models Planned
Soft clustering via EM, BIC to pick the count, and novelty detection.
- E6
A labelling workflow Planned
Read raw data, map, guide, seed, scale, audit, lock the test set.