public void testBasic() {
learner1.setExperimentation(0.99);
System.out.println("testBasic()");
SummaryStats stats = new SummaryStats("action");
int correctActions = 0;
for (int i = 0; i < 100; i++) {
int action = learner1.act();
stats.newData(action);
if (action == CORRECT_ACTION) {
learner1.reward(1.0);
correctActions++;
} else {
learner1.reward(0);
}
checkProbabilities(learner1);
}
System.out.println("final state of learner1 = " + learner1);
System.out.println("learner1 score = " + correctActions + "%");
System.out.println(stats);
}
public void testDistribution() {
System.out.println("\ntestDistribution()");
double q[] = {55, 5, 5, 5, 5, 5, 5, 5, 5, 5};
SummaryStats action1Data = new SummaryStats("action1");
for (int r = 0; r < 10000; r++) {
learner1 = new NPTRothErevLearner(10, 0.2, 0.2, 1, prng);
learner1.setPropensities(q);
SummaryStats choiceData = new SummaryStats("choice");
int action1Chosen = 0;
for (int i = 0; i < 100; i++) {
int choice = learner1.act();
choiceData.newData(choice);
action1Chosen = 0;
if (choice == 0) {
action1Chosen = 1;
}
action1Data.newData(action1Chosen);
}
}
System.out.println(action1Data);
assertTrue(action1Data.getMean() <= 0.57 && action1Data.getMean() >= 0.53);
}
