Udacity Machine Learning NanoDegree Capstone Project

Jeff Gerlach
25 July 2019

Toxic Comment Classification Challenge

Identify and classify toxic online comments

https://www.kaggle.com/c/jigsaw-toxic-comment-classification-challenge

Project Setup

This machine learning project uses relatively smaller datasets, so development was performed locally using a Jupyter .ipynb notebook on a Conda environment.

Jupyter Notebook file:

comment_toxicity-final.ipynb

Python libraries required:

• numpy
• pandas
• matplotlib.pyplot
• IPython.display
• seaborn
• sklearn
• scipy
• keras
• skmultilearn
• scikitplot

Data files:

Kaggle training and test data can be found here - (sample_submission.csv, test_labels.csv, test.csv, train.csv)

fastText word embedding can be found here - (crawl-300d-2M.vec)