Problem 3

Adaboost performance

Train error rate 0

Test error rate 0.058

Negative examples

Positive examples

Experiment log

-- ARGUMENTS ---------------------------------------------------------
./vision_test \
  --problem_number=3 \
  --nb_train_samples=1000 \
  --result_path=./results_problem_3/ \
  --progress_bar=no \
  randomize-train \
  adaboost \
  compute-train-error \
  compute-test-error
-- PARAMETERS --------------------------------------------------------
   "niceness" "15"
   "random_seed" "0"
 * "result_path" "./results_problem_3/"
   "classifier_name" "default.clf"
 * "progress_bar" "no"
 * "problem_number" "3"
 * "nb_train_samples" "1000"
   "nb_test_samples" "1000"
   "nb_weak_learners" "1000"
   "nb_optimization_weak_learners" "100"
   "nb_sampled_samples" "-1"
-- COMPUTATIONS ------------------------------------------------------
Generating the training set.
Building and training adaboost classifier.
Final loss is 0.00194826
TRAIN_ERROR_RATE ADABOOST 0
TEST_ERROR_RATE ADABOOST 0.058
-- FINISHED ----------------------------------------------------------