Duration
2 days
Training Structure
Joint Continuing Education Office, Faculty of Sciences
Overview
This course covers the main steps in supervised classification—from data preparation to training, evaluating results, and deploying a model—using examples based on tabular and text data. It relies on common Python tools and also incorporates the use of language models for text processing. It thus covers both traditional machine learning methods and methods currently used for text data.
The Benefits of the Training Program
Next training session: Spring 2027
Objectives
- Understand the process of a supervised classification project, from data preparation to model deployment.
- Implement reproducible pipelines with Scikit-Learn on tabular and text data.
- Compare a traditional approach to text classification with an initial approach based on a pre-trained language model (BERT).
Program
14 hours of in-person training - Training available for in-house sessions
DAY 1: TABULAR DATA: BUILDING A ROBUST CLASSIFIER WITH SCIKIT-LEARN (7 HOURS)
DAY 2: TEXTUAL DATA: TRADITIONAL METHODS, FOLLOWED BY AN INTRODUCTION TO PRE-TRAINED LANGUAGE MODELS (7 HOURS)
Admission
Target Audience
Professionals who wish to learn a clear and reproducible approach to developing, evaluating, and comparing supervised classification models on tabular and text data.
Tuition Fees
Training fee: €1,190 (including tax)
Mandatory Prerequisites
- Python (basics) and using notebooks
- Basic Data Manipulation (Reading Files, Tables)
- A Google account to run the assignments on Google Colab