Toxic Comments Kaggle Challenge

Developer documentation

1. Python version: 3.7.0

2. Requirements

PIP packages

• numpy
• tensorflow
• keras
• pandas
• nltk
• matplotlib
• scikit
• xgboost
• seaborn
• wordcloud
• PIL
• spacy

To install a package X, run python3 -m pip install X

Jupyter

To install, refer: https://jupyter.readthedocs.io/en/latest/install.html

3. NLTK Corpus

To install the NLTK corpus,

import nltk
nltk.download()

The following are required to run the notebooks:

• Stopwords Corpus

• WordNet Corpus

4. Notebooks to run

1. Data Visualization
2. Data Preprocessing for Decision Tree, Logistic Regression and Ensemble methods: Preprocessing-with -NLTK.ipynb
3. Data-Decision Tree: Decision-Tree.ipynb
4. Logsitic Regression: Logsitic-Regression.ipynb
5. Ensemble: Voting-ensemble.ipynb
6. XGBoost: XGBoost_with_Processed_data_Latest.ipynb
7. Data Preprocessing for LSTM: Preprocessing-with-Dataframe.ipynb
8. LSTM: LSTM_Empiricial_Analysis.ipynb