More than One Way to Count

Our Virtual Fortnightly Machine Learning Meeting Notes for March 10, 2022

and

Mar 22, 2022

Hey Outliers,

Last town hall we had a pair programming session for Chapter 4 of Feature Engineering for Machine Learning. Chapter 4, The Effects of Feature Scaling: From Bags-of-Words to TF-Idf, was focusing on creating features from the English language. Lisa created a notebook that complemented the chapter using restaurant reviews. She created features using Bag-of-Words (BOW) and Term Frequency Inverse Document Frequency (TF-IDF). Then Lisa applied both a Naive Bayes Classifier and a Logistic Regression to try to predict if the review’s sentiment was positive or negative. In addition she compared which feature engineering (BOW, TF-IDF, and L2) provided the best results.

This was our first time we attempted a code-review during a town hall. Lisa was speaking but I was sharing the screen and her code. As we went through the code, we asked her questions why she did certain things. This format was a success. It felt more engaging and we will add this to our arsenal of methods to help each other become better data scientists.

Also, we are changing the schedule of our town halls. Originally we had it every other week but it might be better to change it to once a month. BUT we are in the talks of having a meet just for Kaggle. Let us know if that would be interesting. We are looking for people who would like to host these Kaggle meets. They can present their code, do a livestream coding, or even go through published notebooks. If this is you, contact us and we can get this set!