def main_ACTION(): st.SESSION_CUT = 2 print("***Computing training features") st.CASE = 'training' rd.process_files(st.CASE) print('***Evaluating on the test set') st.CASE = 'test' rd.process_files(st.CASE) return
def main_ACTION_SEQUENCE(): st.SESSION_CUT = 1 st.CASE = 'training' print("***Computing training features") rd.process_files(st.CASE) print('***Evaluating on the test set') st.CASE = 'test' rd.process_files(st.CASE) tc.NUM_NEGATIVE_SAMPLES_PER_CLASS = 70 tc.NUM_POSITIVE_SAMPLES = 630 # in case of this modeling unit (sequence of actions), from each test session we extract exactly one feature vector tc.evaluate_test_actions2(st.TRAINING_FEATURE_FILENAME, st.TEST_FEATURE_FILENAME) return
def main_ACTION(): st.SESSION_CUT = 2 print("***Computing training features") st.CASE = 'training' rd.process_files( st.CASE ) print('***Evaluating on the test set') st.CASE = 'test' rd.process_files(st.CASE) print('EVAL_TEST_UNIT: '+str(st.EVAL_TEST_UNIT)) tc.NUM_NEGATIVE_SAMPLES_PER_CLASS = 200 tc.NUM_POSITIVE_SAMPLES = 1800 if st.EVAL_TEST_UNIT == 0: # evaluation: all actions from a session tc.evaluate_test_session(st.TRAINING_FEATURE_FILENAME, st.TEST_FEATURE_FILENAME) else: # evaluation: action by action tc.evaluate_test_actions(st.TRAINING_FEATURE_FILENAME, st.TEST_FEATURE_FILENAME, st.NUM_EVAL_ACTIONS) return