Predicting Tumor Malignancy With a Decision Tree Classifier

A decision tree created using the ID3 training algorithm.

Data Set: Wisconsin Breast Cancer Diagnostic Data Data Set URL: http://archive.ics.uci.edu/ml/machine-learning-databases/breast-cancer-wisconsin/wdbc.data

What's a Decision Tree?

A decision tree is a tree-based predictive model that makes a guess as to the value of some target variable based on several input variables. If this target variable is a real number then the decision tree is known as a regression tree, while one created to predict a discrete target variable is termed a classification tree. As you can see from the illustration above of a decision tree that I pulled off of google images, a series of decisions are made that take one from the root down to a leaf node, where a determination as to the value of the target variable can be made.

This article has more information on the theory behind decision trees learning. All the code for the decision tree classifier is available via this jar file and can be used with any dataset that the user wishes.

Okay, I know what a decision tree is, but what does it have to do with this project?

I created a decision tree using this breast cancer data set describing physical characteristics of both benign and malignant tumors.

How to run:

First compile using:
ant -buildfile ant -buildfile api/DecisionTreeClassifier/decisionTreeClassifier/src/build.xml all (run from directory containing api directory)

Then start server using: ant -buildfile api/DecisionTreeClassifier/decisionTreeClassifier/src/build.xml all -Darg0 (run from directory containing api directory)

You might need sudo