def __store_stats(self):
learner_name = TornadoDic.get_short_names(self.learner.LEARNER_NAME)
detector_name = self.drift_detector.DETECTOR_NAME
detector_setting = self.drift_detector.get_settings()
file_name = learner_name + "_" + detector_name + "." + detector_setting[
0]
st_wr = open(self.__project_path + file_name.lower() + ".txt", "w")
lrn_error_rate = PredictionEvaluator.calculate_error_rate(
self.learner.get_global_confusion_matrix())
dl, tp, fp, fn = DriftDetectionEvaluator.calculate_dl_tp_fp_fn(
self.__located_drift_points, self.__actual_drift_points,
self.__drift_acceptance_interval)
if len(self.__located_drift_points) != 0:
# learner stats
lrn_mem = numpy.mean(self.__learner_memory_usage)
lrn_ave_runtime = numpy.mean(self.__learner_runtime)
lrn_total_runtime = self.learner.get_total_running_time()
# ddm stats
ddm_mem = numpy.mean(self.__drift_detection_memory_usage)
ddm_avg_runtime = numpy.mean(self.__drift_detection_runtime)
ddm_total_runtime = self.drift_detector.TOTAL_RUNTIME
else:
lrn_mem = asizeof.asizeof(self.learner, limit=20)
lrn_ave_runtime = self.learner.get_total_running_time()
lrn_total_runtime = lrn_ave_runtime
ddm_mem = asizeof.asizeof(self.drift_detector, limit=20)
ddm_avg_runtime = self.drift_detector.TOTAL_RUNTIME
ddm_total_runtime = ddm_avg_runtime
stats = learner_name + " + " + detector_name + ": " + "\n\t" + \
"Classifier Error-rate: " + "%0.2f" % (100 * lrn_error_rate) + "%" + "," + "\n\t" + \
"Classifier Average Memory Usage (bytes): " + "%0.2f" % lrn_mem + "," + "\n\t" + \
"Classifier Average Runtime (ms): " + "%0.2f" % lrn_ave_runtime + "," + "\n\t" + \
"Classifier Total Runtime (ms): " + "%0.2f" % lrn_total_runtime + "," + "\n\t" + \
"Detection Delay: " + "%0.2f" % dl + " TP: " + str(tp) + " FP: " + str(fp) + " FN: " + str(fn) + "," + "\n\t" + \
"Average Detection Memory Usage (bytes): " + "%0.2f" % ddm_mem + "," + "\n\t" + \
"Average Detection Runtime (ms): " + "%0.2f" % ddm_avg_runtime + "," + "\n\t" + \
"Total Detection Runtime (ms): " + "%0.2f" % ddm_total_runtime + "," + "\n\t" + \
"Drift Points detected: " + str(self.__located_drift_points)
print(stats)
st_wr.write(stats)
st_wr.close()
def __store_stats(self):
st_wr = open(self.__project_path + TornadoDic.get_short_names(self.learner.LEARNER_NAME).lower() + ".txt", "w")
lrn_error_rate = PredictionEvaluator.calculate_error_rate(self.learner.get_global_confusion_matrix())
lrn_mem = asizeof.asizeof(self.learner, limit=20)
lrn_runtime = self.learner.get_total_running_time()
stats = self.learner.LEARNER_NAME + "\n\t" + \
"Classifier Error-rate: " + "%0.2f" % (100 * lrn_error_rate) + "%" + "\n\t" + \
"Classifier Memory Usage (bytes): " + "%0.2f" % lrn_mem + "\n\t" + \
"Classifier Runtime (ms): " + "%0.2f" % lrn_runtime
print(stats)
st_wr.write(stats)
st_wr.close()
