Foundations

Getting started with machine learning

The Foundations Workstream aims to provide a path to gaining competency in common statistical and machine learning techniques by:

• creating a roadmap of methods to learn
• gathering together relevant learning materials and,
• developing notebooks in R and Python with example code, where the methods are applied to reserving data sets.

Learning resources

Some learning resources are listed below.

Working party blog articles

See the bottom of this page for a list of all our Foundations blog posts. The entry point is our introduction post which includes a suggested roadmap to learning machine learning.

Other IFoA curated resources

Name	Format	Content	Level	Language	Review
IFoA Data Science Working Party	Website	Resources curated by the IFoA Data Science Working Party	Mixed	R / Python	Useful complementary resource to our working party
IFoA Data Science Working Party articles	Website	List of articles by the working party on data science topics	Mixed	R / Python	Includes a helpful worked example on XGBoost
IFoA Data Science page	Website	List and links of qualifications in data science, courses, IFoA webinars and other tips on how to get started in data science	Mixed	R / Python	Learning about data science can be overwhelming due to the abundance of resources available. It is helpful to know what the IFoA recommends
IFoA Certificate in Data Science	Website	See website	Beginner	Python	Review to be provided in 2021 after completion of the course by a member of our working party

Online courses

Name	Format	Content	Level	Language	Review
Stanford Statistical Learning	Course on videos with R examples via website or EDX app	Mostly supervised learning but also touches on unsupervised learning	Beginner	R (but see note below for python)	Very helpful, the course serves as a good reminder on specific statistical techniques while also making the bridge to more advanced techniques used in machine learning
Deep Learning Specialisation on Coursera	Online course via Coursera website or app	Foundations of Deep Learning, case studies, practice of Python and TensorFlow	Beginner	Python	This is a thorough course. It teaches from first principles what neural networks are and how they work. It does require some familiarity with Python or at minimum with coding
TensorFlow in Practice Specialization on Coursera	Online course via Coursera website or app	Neural Networks with a focus on TensorFlow and its applications	Beginner	Python	This is a practical course on Neural Networks. It focuses on applications and less on deep theory
Kaggle	Website	Wide range of content	Mixed	R / Python (micro-courses in ML use python)	The section Courses allows the user to focus their learning on particular areas. The section Notebooks gives access to other users' content that include tutorials, reproducable code and data.

Notes:

1. Stanford EDX course - Github users have created R and python versions of the course. It is also supported by the textbook An Introduction to Statistical Learning (see below).

2. Kaggle - The notebook on XGBoost and course on Deep Learning may be of particular interest

Textbooks

Name	Format	Content	Level	Language	Review
An Introduction to Statistical Learning	Book (freely available)	Textbook supporting the Stanford Statistical Learning Course	Beginner	N/A	This book is regularly referred to in the Stanford Statistical Learning online course and both follow the same structure
The Elements of Statistical Learning	Book (freely available)	More advanced version of the previous textbook	Intermediate	N/A	Provides more details around the different methods discussed in the introductory text book, An Introduction to Statistical Learning
R for Data Science	Book (freely available)	Introductory guide to data science in R	Beginner	R	Easy guide to common data science tasks in R using the tidyverse

Posts

Talks