Course Material for the machine learning in financial context bootcamp

Welcome to MLiFC

Hi! You found the MLiFC GitHub repository. What is MLiFC you ask? MLiFC is short for machine learning in financial context, a new textbook that aims to teach business, economics and social science students practical machine learning and its applications in business and finance. It is written for the MLiFC bootcamp, taught at turing society rotterdam, also known as ‘Bletchley’.

You can apply to the Bletchley bootcamp taught in Rotterdam from February to April 2018 here. For students who have recieved an invitation to complete the Python challenge, you can find it here

The content here is mostly developed by Jannes Klaas with the support of a community of technical and non-technical people. You are welcome to join us.

Table of content

This ToC gets updated as new chapters come online:

• Week 1: Intro to ML, underlying math and building an NN from scratch

  • Ch. 1 - Intro
  • Ch. 2 - Neural Network from scratch (Pt. 1)
  • Ch. 3 - Training process and the learning rate.ipynb
  • Ch. 4 - A deeper network & Overfitting
  • Ch. 5 - Multiclass Regression
  • Ch. 6 - Minibatch Gradient Descent

• Week 2: Structured data

  • Ch. 7 - Intro to Keras
  • Ch. 8 - The machine learning workflow
  • Ch. 9 - Feature engineering & data preparation
  • Ch. 10 - New frills for our neural network
  • Ch. 11 - Evaluating and deploying the model
  • Ch. 12 - Trees and Ensembles

• Week 3: Computer vision and convolution

  • Ch. 13 - Intro to Computer Vision
  • Ch. 14 - Building an Image classifier
  • Ch. 15 - Theory Interlude - Tensor and Flow
  • Ch. 16 Ch. 16 - Yolo object localization

• Week 4: Sequence models and natural language processing

  • Ch. 17 - NLP and Word Embeddings
  • Ch. 18 - Temporal Order Matters
  • Ch. 19 - LSTM for Email classification

• Week 5: Debugging

• Week 6: Reinforcement learning & the road to AI

• Week 7: Roundup & Final project

It is intended as a jumping board into ML & AI, a basis from which students can dive deeper into specific topics that interest them.

What is this repository?

In this repository we are developing all the course material, free and open for everyone. There is not much good material to help business folk learn about machine learning. So we set out to create our own! The bootcamp that we are developing this material for aims to teach business and finance majors the basics of ML, and especially deep learning, over the course of eight weeks. The theoretical content is not disseminated by lecture, but through iPython notebooks. They are interactive and students can immediately mess around with the code examples provided. The heart of the course are weekly challenges, much like kaggle competitions. Here students gain a lot of practical understanding. As you might have noticed, the content in this repository is unfinished. This is on purpose, as we want to develop the content in the open and under the scrutiny of the community. We believe we can create better content that way.

I found a mistake / something is not clear, what do I do?

File an issue! We love people finding mistakes, so if you find an issue please let us know by filing an issue on GitHub. Even better, if you know a fix, fix it and send a pull request!

Can I use this for my classes?

Yes of course! The content is licensed under an MIT license, meaning you can use it for your class as well. We’d love to hear from you and share knowledge, so just shoot an email to jannes [at] tsociety.io to get in touch.