Weekly Readings 24th-30th/May/2021
Interesting news/blogs that I read during a week.
1. DataPrep v0.3.0 has been released
EDA is not an easy stuff for me because I’m not type of person who enjoys going to details. That’s why I’ll search for EDA python libraries which help me do EDA. Before knowing about DataPrep, I knew Sweetviz. You can see there are other python libraries can help you doing EDA in the link below:
4 Libraries that can perform EDA in one line of python code
Exploratory data analysis using Pandas Profiling, Sweetviz, Autoviz, and D-Tale
To be honest, I’m not happy with Sweetviz because I love something simple and easy-to-use. So I’m quite exciting when I first see DataPrep, because it’s maybe the right thing that I want to deal with EDA and to use for data preprocessing.
Hope that you can utilize this tool for your work too.
2. Getting Started With Testing in Python
Getting Started With Testing in Python - Real Python
In this in-depth tutorial, you'll see how to create Python unit tests, execute them, and find the bugs before your…
In a blog, I said that I should improve the quality of codes by writing unit tests for them. The article above shows how I create an unit test.
It’s not necessary to be detailed, perfect, just write tests that you can do for your code. It’s good enough for starters.
3. Multi-Class classification using Focal Loss and LightGBM
Multi-Class classification using Focal Loss and LightGBM
Using Focal Loss for multi-class classification
An excellent blog! The author explained his thinking approach and the solution very well. Deal with imbalanced data is a big problem in academic world and practical world also. In current research articles that I have read, using Focal Loss is a new promising solution for this problem. I have seen people utilize Focal Loss for solving binary classification problems, and this is the first time I see someone use it to dealt with multi-classification problems.
I hope that I can apply this article to the problems that I solve in real life. Now, it’s a big kudos for the author.
4. Other good articles:
Best Tools for Model Tuning and Hyperparameter Optimization - neptune.ai
I vividly remember a machine learning hackathon that I participated in two years ago, when I was at the beginning of my…
ML Model Interpretation Tools - What, Why, and How to Interpret - neptune.ai
Interpretation is literally defined as explaining or showing your own understanding of something. When you create an ML…