Skip to content
racousinPublic

About

No description, website, or topics provided.

Resources

Stars

6 stars

Watchers

0 watching

Forks

Latest commit

 

History

27 Commits

Folders and files

Repository files navigation

L2 Data Science & Text Mining — Practicals

Course for L2 Math students 6 sessions + 1 project.

Structure

Session Topic Notebook
Session 1 Foundations of ML Open In Colab
Session 2 Non-Parametric & Ensemble Models TP1 (Regression): Open In Colab
Session 3 Ensemble Methods Open In Colab
Session 4 Neural Networks & Tuning Open In Colab
Session 5 Text Extraction & Classification Open In Colab
Session 6 Embeddings & Word Representations Open In Colab
Project Temperature Prediction (24h) Open In Colab

Content Summary

Session 1 — Foundations of ML

Linear regression, logistic regression, MSE/MAE/R², precision/recall/F1/AUC, preprocessing (scaling, encoding, imputation), train/test split, k-fold cross-validation, sklearn pipelines.

Session 2 — Non-Parametric & Ensemble Models

Theory notebooks (tp_0): KNN (distance metrics, effect of k, validation curve, curse of dimensionality), Decision Trees (Gini/entropy, depth control, feature importance), SVM (maximum margin, soft margin C, kernel trick, RBF gamma), Bagging & Random Forests (bootstrap, variance reduction, OOB estimation, feature randomization), Gradient Boosting (AdaBoost, residual fitting, learning rate), Regularization (Ridge L2, Lasso L1 with feature selection, Elastic Net). TP1 — Regression: Model comparison on Diabetes dataset (baseline → KNN, Tree, SVR, RF, GB → GridSearchCV → cross-validation). TP2 — Classification: Model comparison on Wine dataset (scaling demo, baseline → 5 models → GridSearchCV on SVM → cross-validation).

Session 3 — Ensemble Methods

Bias-variance tradeoff (empirical), Random Forest (bagging, n_estimators, feature importance), Gradient Boosting (learning rate, sequential correction), full model comparison with CV.

Session 4 — Neural Networks & Tuning

MLP with sklearn (architectures, activations), loss curves, GridSearchCV, RandomizedSearchCV, comparison with ensemble methods.

Session 5 — Text Extraction & Classification

Tokenization, stopwords, Bag of Words, TF-IDF, text classification pipeline on 20 Newsgroups, model comparison (LogReg, NB, SVM), feature importance for text.

Session 6 — Embeddings

Pretrained Word2Vec exploration (analogies, similarity), document embeddings (mean pooling), embeddings vs TF-IDF comparison, conceptual intro to transformers.

Project — Temperature Prediction

Predict 24h temperature for 10 cities from 1000 cities x 2 weeks hourly data. Teams of 2-3, evaluated on ML-Arena. Grading: 50% performance + 50% presentation.

Requirements

numpy
pandas
matplotlib
seaborn
scikit-learn

About

No description, website, or topics provided.

Resources

Stars

6 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages