Wei-Hung Weng (MIT)
We would like to introduce basic machine learning techniques and toolkits for clinical knowledge discovery in the workshop. The material will cover common useful algorithms for clinical prediction tasks, as well as the diagnostic workflow of applying machine learning to real-world problems. We will use Google colab / python jupyter notebook and two datasets:
- Breast Cancer Wisconsin (Diagnostic) Database, and
- pre-extracted ICU data from PhysioNet Database
to build predictive models.
The learning objectives of this workshop tutorial are:
- Learn how to use Google colab / jupyter notebook
- Learn how to build machine learning models for clinical classification and/or clustering tasks
To accelerate the progress without obstacles, we hope that the readers fulfill the following prerequisites:
- [Skillset] basic python syntax
- [Requirements] Google account OR anaconda
The repository include three tutorial jupyter notebooks. You may want to download it or run it on colab.
- In part 1, we will go through the basic of machine learning for classification problems. The tutorial is modified from the version provided by Dr. Alistair Johnson. (Thanks!)
- In part 2, we will investigate more on unsupervised learning methods for clustering and visualization.
- In part 3, we will play with simple neural networks.