Pensez comme un Data Scientist

• La formation se compose de travaux pratiques (mises en situation, débats, exercices).
• Une évaluation quotidienne de l'acquisition des connaissances de la veille est effectuée.
• Une synthèse est proposée en fin de formation.
• Une évaluation à chaud sera proposée au stagiaire à la fin du cours.
• Un support de cours (version électronique) sera remis à chaque participant comprenant les slides sur la théorie, les exercices.
• Une feuille d'émargement par demi-journée de présence est fournie en fin de formation ainsi qu'une attestation de fin de formation si le stagiaire a bien assisté à la totalité de la session.

Cette formation vous initie à la Data Science et au Machine Learning. Vous apprendrez à penser comme un Data Scientist, de la compréhension des données à la mise en œuvre d'algorithmes de Machine Learning avec Python.

• Comprendre les fondamentaux du Machine Learning
• Maîtriser les bases du langage Python pour la Data Science
• Savoir choisir les algorithmes adaptés à un problème
• Mettre en pratique les concepts sur des jeux de données réels

• Architectes
• Chefs de projet
• Développeurs

Le formateur est un expert du domaine qui intervient sur le sujet depuis plusieurs années en formation mais aussi en conseil.

• Aucun prérequis technique spécifique, cette formation est accessible à tous.

The basics of machine learning

• A few examples by way of introduction
• Generalization: a founding principle of Machine Learning to enable prediction and segmentation
• Supervised and unsupervised algorithms
• Appropriation of the Data Scientist vocabulary

First steps

• Collecting and storing data
• Analyze, understand, clean and structure data: Feature Engineering
• Learning and creating a model
• Evaluating the model
• Improving the model

The Python language

• The syntax of the language
• The development tools: Jupyter notebook
• Data scientist libraries: Pandas, Scikit-learn
• Analyze and understand data
• Matplotlib and Seaborn: Data visualization libraries for Python

Choosing machine learning algorithms

• Understanding the issues
• Overview of the main algorithms
• Classification: k-Nearest Neighbors (k-NN),
• Decision Tree, Random Forest, XGBoost
• Regression: Logistic regression
• Clustering: K-Means, DBScan
• The different scoring methods

Advanced concepts

• Cross Validation
• Machine Learning Ensemble: combine algorithms for better accuracy
• Automate data manipulation with a pipeline

Process data in parallel

• Why parallelize?
• Adapting algorithms
• Complementary complexity
• Distribution frameworks available: Spark and Dask

Deploying in production

• Integrating a pipeline into an automated deployment chain (continuous delivery)
• Packing a model: Predictive Model Markup Language
• Creating a REST endpoint with Flask (Python)
• Deploying in the cloud

Deep Learning

• General presentation: neural networks
• Convolutional neural networks: Image analysis
• Recurrent neural networks
• Understanding by example: Pytorch & Keras

Training must serve the performance of the employee and the company

This requires a constant quest for excellence on the part of the training organization with a systematic adaptation to the company's challenges, the regular updating of course materials, and a technological watch that is essential to always be at the cutting edge of the field.

Technical expertise is as important as teaching skills

• Our trainers are all experts in their field. But what do they have that others don't? We select them for their teaching skills and their teaching methods. We place pedagogical qualities on the same level as expertise so that our trainees get the most out of their training.
• Excellence is born of excellence
• Many of our clients are among the leaders in their respective industries or the most promising start-ups. We know that training the employees of such companies requires attention to every detail and coaching that matches the ambition of our trainees. That's why we know how to turn today's leaders into tomorrow's champions!

Nous consulter