Mushroom-Classifier

Contains a neural network powered binary-classifier for the UC Irvine Mushroom Dataset.

Table of Contents generated with DocToc

• Dependencies
• Usage Instructions
  • Install dependencies
  • Get data set and generate feature vectors
  • Data Set splitting
  • Run and test model
• Data Pre Processing
• Neural Network Specifications
• Classifier Performance

Dependencies

• Python 2.7
• tensorflow
• tflearn
• hickle

Usage Instructions

Install dependencies

Run source setup.sh or ./setup.sh
In case you run into issues with permissions, run sudo chmod +x setup.sh and try running the file again.

Get data set and generate feature vectors

Run generate_datasets.py

Data Set splitting

This stage involves dividing the data set into three parts:

• Training
• Validation
• Testing

By default, they are split in the ratio 9:1:1 (Training: Validation : Testing). This ratio can be modified by changing the TRAIN, VALID and TEST constants in split_datasets.py When ready, run split_datasets.py

Run and test model

run nn_model.py

Data Pre Processing

For each feature, attributes are one-hot encoded. Missing values are represented as an independent bit in the one-hot encoded representation. These encoded attributes are then chained together to form a 126-bit long feature vector.

Neural Network Specifications

This binary classifier uses one hidden layer in addition to an input and output layer.
The particulars of each layer are described as under:

• Input Layer (126 nodes)
• Hidden Layer (64 nodes; activation function : relu)
• Output Layer (2 nodes; activation function : softmax)

Classifier Performance

On training the classifier on 90 % of the data (80% training + 10% validation), this model has achieved an accuracy of 100% on unseen Test data! Yay!