Kaggle Results - EDA and Machine Learning

My Kaggle competition notebooks and submissions.

This collection of Kaggle competition notebooks demonstrates my ability to:

1. Explore and visualize the data using plotting libraries, statistics and custom functions.
2. Load, clean, and prepare the data for modeling.
3. Train the model (choosing an algorithm, metric, train/test/holdout). Cross-validation and iterating over model params to combat bias/variance in model.
4. Using the model to predict on holdout/unseen data.

The types of models seen in the notebooks vary from:

Natural Language Processing and the use of Keras LSTM networks, decision trees and linear models. Computer vision using Keras custom CNN networks for multi-class classification. Time-series regression modeling over years of data using custom forward chaining validation technique with xgboost. Binary classification predictions.