FastAI2022

Lesson 5

Lesson 5 did not contain too many new topics for me, but instead of moving on fast, I resisted the temptation and did a bit of recapping.

I returned to the titanic competition for working on EDA, feature engineering and also building a Fast.AI model. I explored how I could co-work with ChatGPT on this task. As a result I created the following notebooks:

• Titanic EDA and 1R with ChatGPT): The goal of this notebook is to create a 1R-model (One Rule, also see related paper) of the Titanic Kaggle Competition after doing EDA (Exploratory data analysis), supported by ChatGPT.
• Titanic Feature Engineering and Model Creation with ChatGPT): The goal of this notebook is do some feature engineering on the dataset of the Titanic Kaggle Competition and to later create a Fast.AI-learner to submit the result to Kaggle, supported by ChatGPT.

I also wrote about my thoughts on working with ChatGPT and what I have learned along the way in this blog post.

Still on the bucket list for further research / things to be done while moving on:

• Transfer to Spaceship Titanic including Feature engineering
• Watch lesson 4 + 6 and 7 of the 2020 version (which is the book)

This site is open source. Improve this page.