Learning Algorithms

CSI4106 Introduction to Artificial Intelligence

Prepare

• CSI4106 Jupyter Notebooks and Google Colab Tutorial

Participate

• slides (PDF, Jupyter Notebook)

Practice

Classifiers from first principles

• Building a Decision Tree Classifier (Jupyter Notebook, Open in Colab)
• Building a K-Nearest Neighbours Classifier (Jupyter Notebook, Open in Colab)

Python and datasets

For those who are not yet familiar with numpy and pandas, it is recommended to explore these tutorials over the coming weeks at your own pace. Aurélien Géron’s notebooks offer insights into essential machine learning concepts.

• numpy – a fundamental library centered around N-dimensional array objects.
• pandas – powerful data analysis tools centered around the DataFrame.
• visualization – demonstrates how to use matplotlib to produce beautiful graphs.

In the upcoming weeks, you may wish to apply the concepts discussed in class to a variety of datasets to further develop your skills and intuition. Below is a list of websites where you can find relevant datasets.

Small-to-medium datasets, very student-friendly

• UCI Machine Learning Repository
• scikit-learn toy datasets

Larger datasets

• OpenML, integrates directly with scikit-learn
• PMLB (Penn Machine Learning Benchmarks)

Competition(s)

• Kaggle Datasets
  • Kaggle, a platform owned by Google, serves as an online community tailored for data scientists and machine learning practitioners. It facilitates participation in data science competitions, collaboration on various projects, and provides access to diverse datasets. Additionally, users can build models using its web-based tools.