lkurlandski/Decision-Tree-Learner

Decision-Tree-Learner

CSC 426 Project 3

Authors:

Lana Abdelmohsen, Luke Kurlandski, Gordon Petry, Jason Swick

Description:

A basic descision tree learning algorithm, capable of learning a tree from multiple data sources and producing a visual output.

Content Guide and Description of Code

1. main.py

• primarily for executing the code
• contains a function to process input files
  • reads the csv/txt file, learns the attributes from the first row, the target concept from the final column, ignores column named "ExampleID"
  • processes each successive row into a tuple, where the first element is a dict of attribute : value pairs and the second element is the boolean target concept

2. decision_tree.py

• contains the DecisionTree class, which
  • contains ID3 algorithm
  • computes information gain
  • computes entropy of a set
  • converts itself to a string representation
  • prints itself to a file
• contains the Node class, which
  • is a simple, non-optimized n-tree with labeled edges
  • serves as the backbone for the DecisionTree class
  • can convert itself into a string representation

3. D2_output.txt

• contains string representation of the decision tree as it is being built for the PlayTennis task

4. D3_output.txt

• contains string representation of the decision tree as it is being built for the EnjoySport task

5. D3.pdf

• a repsonse to the version space question

6. D4_input.csv

• contains our own input data for PlayTennis task for task 4

7. D4_output.txt

• contains our own output data for PlayTennis task for task 4

8. D5.pdf

• group deliverable

9. other .txt, .csv

• other files such as the project description etc.

Instructions for Use

1. Compatibile with python 3.6.0.
2. TCNJ HPC specific instructions:

• module add python/3.6.0

3. To run the program:

• python3 main.py input_path output_path

4. where

• input_path : name of a .csv file that contains training data
  • may also be .txt file type, but must contain comma separated values
  • last column should be value of target attribute
• output_path : name of a .txt file to write the learned decision tree
  • NOTE: the decision tree is printed sideways, not vertically!

Output

A sideways-printed decision tree will be output the the output_path supplied when running the program.