{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Intro\n", "\n", "Andreas Mueller\n", "NYU Center for Data Science\n", "\n", "## Clone this repository:\n", "https://github.com/amueller/advanced_training\n", "\n", "## Schedule\n", "\n", "Start: 9am\n", "End: 5pm\n", "\n", "Lunch: ~11:45am\n", "Snack: ~3pm" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Outline\n", "\n", "## 1 Basic algorithms\n", "- Review of supervised learning\n", "- Linear models for classification and regression\n", "- Loss functions, regularization, empirical risk minimization\n", "- Path algorithms\n", "\n", "## 2 Preprocessing\n", "- Scaling and normalization\n", "- Continuous and discrete features\n", "- Feature selection\n", " - Univariate\n", " - Model-based\n", " - RFE\n", " - Forward / backward selection\n", "- Polynomial and interaction features\n", "\n", "\n", "## 3 Basic tools\n", "- Cross-validation vs train/test split\n", "- GridSearchCV\n", "- Overfitting Parameters\n", "- Scoring Metrics\n", "\n", "## 4 Advanced Supervised Learning\n", "- Decision Tree Recap\n", "- Random Forests\n", "- Gradient Boosting / xgboost\n", "- Kernel SVMs\n", "- Kernel approximation\n", "- Neural Networks\n", "\n", "## 5 Advanced tools\n", "- Pipelines\n", "- FeatureUnion\n", "- Function Transformer\n", "\n", "\n", "## 6 Unsupervised feature extraction and visualization\n", "- PCA\n", "- NMF\n", "- Robust PCA\n", "- TSNE\n", "\n", "\n", "## 7 Outlier Detection\n", "- Elliptic Envelope\n", "- IForest\n", "- KDE\n", "- robust PCA\n", "\n", "## 8 Gaussian Processes\n", "\n", "\n", "## 9 beyond standard sklearn\n", "- out of core\n", "- custom estimators\n", "\n" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.5.1" } }, "nbformat": 4, "nbformat_minor": 0 }