Who is this training for?
Beginner to intermediate developers wishing to discover machine learning Data analysts wishing to integrate predictive techniques into their analyses IT professionals wishing to learn about applied artificial intelligence Students or professionals retraining for data professions Anyone with a basic background in programming and wishing to explore machine learning with Python
Training objectives
- Understand machine learning and its subfields
- Familiarize yourself with some common machine learning algorithms
- Choose the right algorithm to use for your case
- Acquire the expertise to analyze the results of machine learning algorithms
Summary
This course provides a structured and accessible introduction to the fundamental concepts of machine learning using Python. Participants learn the basics of machine learning, as well as the main techniques for preparing, analyzing and modeling data.
Course outline
Introduction:
- Supervised, Unsupervised and Reinforcement Learning
- Classification, Regression, Structure Prediction
- Model Evaluation: Metrics • Hyperparameter and Model Selection
- Introduction to Scikit Learn
- Data Types and Method
Selection Guide Classification:
Introduction with OCR:
- K Nearest Neighbor Algorithm
- Decision Trees and Visualization
- Set Method • Support Vector Machines (SVM)
- Visualization of Results Classification:
Advanced concepts with sentiment analysis:
- Data pre-processing for training
- Dimensionality reduction
- Batch training
- Interpretability (importance of weights, LIME)
Regression:
- Linear regression
- Nonlinear regression with kernel methods
- Outlier data detection and management
- Time series: Challenges, decomposition and prediction methods
- Time series: non-stationary regression and autoregressive models
Recommender System: Case Study:
- Collaborative Filtering by User
- Collaborative Filtering by Article
- Advanced Concepts and Algorithms
Unsupervised Learning:
- Clustering: K-Means, Hierarchical Methods, Density Methods
- Dimensionality Reduction: PCA, t-SNE
- Generative Models:
Introduction to Variational Autoencoders and Autoencoders Debugging How-To:
- Overfitting Testing: Model and Data Size
- Pipeline Testing
- Exploring Alternative Metrics
Approach and methodology
Practical and progressive approach combining focused theory and guided workshops. Participants discover the fundamental concepts of machine learning through concrete exercises in Python, promoting an immediate application of the learning. They experiment with the different stages of a machine learning project: data preparation, algorithm selection, model training, and performance evaluation. The training favours interactivity, experimentation and the resolution of practical cases in order to develop skills that are directly transferable in a professional context.
Prerequisites
Knowing how to program in Python
Recommendations
Review: the basics of Python Familiarize yourself with: data manipulation (tables, lists, files) Identify your objectives: discover machine learning understand algorithms analyze data
