Introduction to applied machine learning - using geotechnical data (Pilot course)
Interested in machine learning and prefer an applied approach? Join our 4-session course and learn machine learning by completing all the typical steps in an end-to-end project using a geotechnical dataset.
Instructor: Tom Frode Hansen
Duration & sessions: 16h course in 4x4h sessions
Description: Interested in machine learning and prefer an applied approach? Join our 4-session course and learn machine learning by completing all the typical steps in an end-to-end project using a geotechnical dataset. You will build a supervised prediction model using a tabular dataset. Basic knowledge of machine learning and Python programming is required. We start by setting up a good project structure and establishing practical tools and practices. The course guides you through data exploration, processing, training, evaluation, and deployment of your model. Instead of coding solutions from scratch, we use well-known, tested, and established solutions. The course offers practical knowledge directly applicable to real machine learning model development. Following scientific principles, we focus on proper configuration, structure, and organisation of variables. Results from experiments are tracked, and promising models are saved. You will apply your new skills through two homework exercises and several smaller tasks. Enrol in our course today and start developing your own machine learning models.
Content: Applied machine learning, implemented using Python 3.12. Organising variables/configuration using Hydra, and experimentation results using mlFlow. The Python packages Scikit-learn, Pandas, Numpy and Matplotlib will be central, with a final deployment of the model in Streamlit. Topics include: quality controlling data, outlier analysis, feature selection, data balancing, metrics, pipelines, hyperparameter optimisation, training and evaluation using cross-validation, deployment of trained model.
Language: English
Required previous experience: Basic knowledge of machine learning and Python programming
Teaching mode: online via MS Teams
Course fee: 8410 NOK (Pilot course with 50% discount: 4210 NOK)
Min-/max. allowed participants: 6/14
Hardware & software requirements: For a smooth learning experience, it's important to have the right tools. You'll need your own personal computer or laptop with an installed IDE, such as VSCode. We'll be using VSCode for teaching, so we highly recommend having it installed. We'll also be running several operations using the terminal installed by default in VSCode. You'll need to have installed the two applications pyenv (to control the Python version), and poetry (for environment handling and Python package control). Use pyenv to install Python version 3.12. Additionally, you'll need Microsoft Teams ready video conference equipment including a microphone and camera. We expect all this to be ready for the first session.
Dates: Pilot course: 5th Nov., 7th Nov., 12th Nov., 14th Nov. (13:00-17:00)
Reservations: NGI's course terms and conditions that apply are given during ticket purchase.
Contact: tom.frode.hansen@ngi.no
