A bachelor's course on the foundations of data mining — how we represent, compare, index, and mine large collections of data. This year the course has an information-access focus, building from distances and similarity through embeddings, matrices, graphs, recommenders, and hashing.
Mondays 13:45 & Thursdays 10:45 · EEMCS Lecture Hall Ampere · labs Tuesdays 08:45 · exam Monday 25 January, 13:30. Teaching team: A. Anand, J. Urbano, N. Tömen, with guest lecturers.
| Week | Monday 13:45 | Thursday 10:45 | Lab · Tue 08:45 | Admin |
|---|---|---|---|---|
| 9 Nov | IntroductionAnand + Urbano | DistancesAnand | — | — |
| 16 Nov | RetrievalUrbano | Matrices & PCATömen | — | — |
| 23 Nov | RankingsUrbano | EvaluationUrbano | — | Assignment published |
| 30 Nov | EmbeddingsAnand | Matrix decompositionTömen | — | Lab announced |
| 7 Dec | No lecture | Recommender systemsTömen | 8 Dec · Implement NMF | — |
| 14 Dec | HashingAnand | Sketchingguest | 15 Dec · Implement MinHash | Practice exam |
| 21 Dec–1 Jan | ❄ Christmas break — no teaching | |||
| 4 Jan | Invited lecture (TBD)guest | IndexingUrbano | — | Assignment due |
| 11 Jan | GraphsAnand | Bias & explanationsguest | 12 Jan · Practice exam | — |
| 18 Jan | Exam preparationTömen | No lecture | — | — |
| 25 Jan | 📝 Data Mining exam · Monday 25 January, 13:30 · Period 2 exam week | |||
What data mining is · DM vs ML vs data management · John Snow & famous examples · a word of caution.
Metrics, cosine / Euclidean / Hamming, and a near-duplicate case study.
From raw data to vectors · word2vec · graph embeddings (DeepWalk / node2vec).
Properties, centrality, Girvan–Newman, spectral clustering, node embeddings.