Simple Sample Codes for Palmer Penguin Dataset Classification

This is a collection of simple and easy-to-read programs for Palmer Penguin classification (for teaching purposes). I use several frameworks to classify this data, so the reader can compare the difference between one and another for the same usage.

This repository is an improvement from the previous iris-python repository. Compared to Iris dataset, Palmer Penguin dataset is little bit challenging, because there are some preprocessing steps that necessary for this data (remove missing values, standardization, etc.).

Each program contains:

1. Load the cleaned data
2. Preprocess (convert to numeric, standardize)
3. Train a model
4. Test

Required Packages:

Please install requirements.txt file in each folder

---

Keras (using Tensorflow backend)

• Simple Neural Net using Keras (Multilayer perceptron model, with one hidden layer)

Python & Scikit-Learn

• Simple Neural Net without external library (Multilayer perceptron model, with one hidden layer)
• Simple Neural Net without external library (No-hidden layer model)
• Simple Neural Net using Scikit-learn MLPClassifier (Multilayer perceptron model, with one hidden layer)

PyTorch

• Simple Neural Net using PyTorch (Multilayer perceptron model, with one hidden layer)
• Simple Neural Net using PyTorch Lightning (Multilayer perceptron model, with one
• Simple Neural Net using Fastai (Multilayer perceptron model, with one hidden layer)

Tensorflow

• Simple Neural Net using Tensorflow (Multilayer perceptron model, with one hidden layer)
• Simple Neural Net using Tensorflow version 2.x (Multilayer perceptron model, with one hidden layer)