Overview

This is referred to as a “Toy Dataset” because of its small size and the fact that it uses made-up data. However, it is meant to show that the probability of someone having cancer increases if they smoke or have friends who smoke.

Target: cancer

The facts contain information on two labels: friends, smokes.

train_facts

friends(Alice, Bob).
friends(Alice, Fred).
friends(Chuck, Bob).
friends(Chuck, Fred).
friends(Dan, Bob).
friends(Earl, Bob).
friends(Bob, Alice).
friends(Fred, Alice).
friends(Bob, Chuck).
friends(Fred, Chuck).
friends(Bob, Dan).
friends(Bob, Earl).
smokes(Alice).
smokes(Chuck).
smokes(Bob).

train_pos

cancer(Alice).
cancer(Bob).
cancer(Chuck).
cancer(Fred).

train_neg

cancer(Dan).
cancer(Earl).

Download

Download: Toy-Cancer.zip (2.83 KB)

  • md5sum:

    fa15b64583f9b1abc7fd78b93025792d

  • sha256sum:

    618d9283caa5459711b01d7b535aa1e91c8c98945ed4085248368a373ce880c2

Setup

  1. After downloading, unzip Toy-Cancer.zip

    unzip Toy-Cancer.zip

  2. If you’re using a jar file, move it into the Toy-Cancer directory:

    mv (BoostSRL jar file) Toy-Cancer/
    mv (auc jar file) Toy-Cancer/

  3. Learning:

    java -jar BoostSRL.jar -l -train train/ -target cancer -trees 10

  4. Inference:

java -jar BoostSRL.jar -i -test test/ -model train/models/ -target cancer -aucJarPath . -trees 10

Modes

useStdLogicVariables: true.
setParam: treeDepth=4.
setParam: nodeSize=2.
setParam: numOfClauses=8.
mode: friends(+Person, -Person).
mode: friends(-Person, +Person).
mode: smokes(+Person).
mode: cancer(+Person).
bridger: friends/2.
//precompute1:
num_of_smoking_friends(x, n) :-
   friends(x, y), // grounding x first
   countUniqueBindings((friends(x,z)^smokes(z)), n).
mode: num_of_smoking_friends(+Person, #Number).

Updated: