heikalb / decision-mushroom

Using decision trees to classify mushrooms as either edible or poisonous

Can we eat this mushroom?

This project is based on a dataset on mushrooms consisting of physical attributes such as cap shape and odor and each mushroom in the sample is classified as either edible or poisonous. The goal here is to train decision trees to predict whether a mushroom is edible or poisonous based on its physical attributes.

decision_tree_scikit.py uses a decision tree from sklearn. decision_tree_scikit.py uses a decision tree from made from scratch. get_mushroom_data.py is used in both of the aforementioned programs to load the mushroom, including extrapolating missing feature values. The decision tree made from scratch achieved 100% accuracy in multiple tests.

In addition to using decision trees, this project has been expanded to include classification using a neural network. This is implemented in classify_nn.py.

Data

Data source: http://archive.ics.uci.edu/ml/datasets/Mushroom

Dataset in this repository: agaricus-lepiota.data

Description of the dataset: agaricus-lepiota.names

Requirements

• scikit-learn
• tensorflow
• pandas