Examples

core.easy_experiment module

core.easy_experiment.easy_experiment(source, target_test, target_retrain=None, domain_adaptive=False, models='all', binary_evaluation=True, evaluation_metrics='all', thread_num=4, remove_time=True, plot=False)

easy_experiment() function allows the user to quickly estimate the occupancy state of a room leveraging the specified models and evaluation metrics. Refer to the sample code for easy_set_experiment() for more details.

core.easy_experiment.easy_set_experiment(source_set, target_test_set=None, split_percentage=0.8, target_retrain=None, domain_adaptive=False, models='all', binary_evaluation=True, evaluation_metrics='all', thread_num=4, remove_time=True, plot=False)

easy_set_experiment() function allows the user to quickly perform occupancy estimation on multiple data sets using the specified models and evaluation metrics. We recommend the use of easy_set_experiment() instead of easy_experiment() to get the formatted return value for Result, which is essential for plotting.

Note

Always pass a list as the argument of load_sample() function when using easy_set_experiment() to obtain a dictionary that maps data set names to Dataset objects.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
import core
import pprint

# Load and separate the train and test data set
data = core.data.load_sample(["umons-all", "aifb-all"])

# Perform occupancy estimation
score, predict_result = core.easy_set_experiment(data, models=["RandomForest"])

pprint(score)
pprint(predict_result)

# Make the scores ready to plot
result = core.evaluation.Result()
result.set_result(score)

# If you somehow want to use "easy_experiment", here is the code
train, test = data["umons-all"].split(0.8)

# Perform occupancy estimation
score, predict_result = core.easy_experiment(train, test, models=["RandomForest"])

pprint(score)
pprint(predict_result)