Data Mining and Social Media

FDMS: "Fouille de Données et Médias Sociaux" course at Université Pierre et Marie Curie.

Practical sessions:

1. Data science tutorial: predict Titanic survivors, see the article on my website.
2. Playing with fishes: polynom interpolation, cross-validation, Akaike information criterion on fish data
3. L1-regularization and parcimonious models
4. Bagging and boosting implementations
5. Data visualization and recommendation: t-SNE and collaborative filtering
6. SoRec: collaborative filtering for movie recommendation
7. Collective inference: classification in graphs, labelling unknown university pages
8. Diffusion: How do diseases/trends propagate?
9. Bandit learning, linear UCB: ad targetting