DAS410108

The course aims to develop a background on models and algorithms based on data for fundamental applications in science and engineering, with a distinct focus on problems found in the identification, modeling, and control of dynamic systems.

Machine Learning Resources for Students

Data Analysis

• scikit-learn - Simple and efficient tools for data mining and data analysis built on NumPy, SciPy, and matplotlib.

• Orange - Visual programming tool for data analysis with interactive data visualization and a component-based approach.

• statsmodels - Package for statistical modeling and hypothesis testing with comprehensive implementation of statistical methods.

• pandas - Data structures and tools for data manipulation and analysis with support for various file formats.

Machine Learning

• scikit-learn - Comprehensive collection of ML algorithms with a consistent interface and model evaluation tools.

• XGBoost - Optimized gradient boosting library that is highly efficient, flexible, and portable.

• CatBoost - Gradient boosting library that handles categorical features automatically with fast training and GPU acceleration.

• LightGBM - Fast, distributed gradient boosting framework with lower memory usage and better accuracy.

• NGBoost - Natural Gradient Boosting for probabilistic prediction with uncertainty estimation.

Feature Engineering

• tsfresh - Automatic extraction of relevant features from time series data with statistical feature generation.

• autofeat - Automated feature engineering and selection that creates non-linear combinations of features.

Hyperparameter Tuning

• Optuna - An open-source hyperparameter optimization framework that automates the search for optimal hyperparameters with efficient search algorithms.

• Hyperopt - Distributed asynchronous hyperparameter optimization library that implements various algorithms for searching hyperparameter spaces.

• Ray Tune - Scalable framework for hyperparameter tuning with support for distributed execution and various search algorithms.

AutoML (Automated Machine Learning)

• auto-sklearn - Automated machine learning toolkit based on scikit-learn with automatic model selection and hyperparameter tuning.

• TPOT - Automated machine learning tool that optimizes ML pipelines using genetic programming.

• H2O AutoML - Automated machine learning with various algorithms and support for Python, R, Java, and Scala.

• PyCaret - Low-code machine learning library that automates the ML workflow and is built on top of several ML libraries.

• FLAML - Fast and Lightweight AutoML library by Microsoft that automatically finds accurate machine learning models with low computational resources.