"maxspeed": "30",
"name": "Heinheimer Stra\u00dfe"
},
"exit_way": {
"highway": "residential",
"maxspeed": "30",
"name": "Vogelsbergstra\u00dfe"
}
}
# Create number of LineString serving as way data
angles = np.linspace(-3*np.pi/8, 11*np.pi/8, 15)
scale = 50.0
coords = [[(0,0),(0,1*scale),(np.cos(a)*scale,(1+np.sin(a))*scale)] for a in angles]
labels = ["Test way %.2f°" % (a/np.pi*180.) for a in angles]
lines = [LineString(c) for c in coords]
test_samples = []
# Create synthetic samples
for line, label in zip(lines, labels):
int_sit, osm = create_intersection_data(line, 0.5, normalized=True, tags=tags)
sample = create_sample(int_sit, osm, label, output="console")
test_samples.append(sample)
train_samples = automatic_test.load_samples('../data/training_data/samples_23_09_15/samples.pickle')
# train_samples = automatic_test.normalize_features(samples)
# Load algorithm with train samples and test the synthetic ones
rf_algo = regressors.RandomForestAlgorithm()
automatic_test.train([rf_algo], train_samples)
results = automatic_test.predict([rf_algo], test_samples)
automatic_test.show_intersection_plot(results, test_samples, orientation="curve-secant")
sys.path.append('../')
import automatic_test
import regressors
import reference_implementations
feature_list = [
"intersection_angle", # Angle between entry and exit way
"maxspeed_entry", # Allowed maximum speed on entry way
"maxspeed_exit", # Allowed maximum speed on exit way
# "lane_distance_entry_exact", # Distance of track line to curve secant center point at 0 degree angle
# "lane_distance_exit_exact", # Distance of track line to curve secant center point at 180 degree angle
"lane_distance_entry_lane_center", # Distance of lane center line to curve secant ceter point at 0 degree angle
"lane_distance_exit_lane_center", # Distance of lane center line to curve secant ceter point at 180 degree angle
"oneway_entry", # Is entry way a oneway street?
"oneway_exit", # Is exit way a oneway street?
"vehicle_speed_entry", # Measured vehicle speed on entry way at INT_DIST
"vehicle_speed_exit" # Measured vehicle speed on exit way at INT_DIST
]
rf_algo = regressors.RandomForestAlgorithm(feature_list)
rfc_algo = regressors.RFClassificationAlgorithm(feature_list, bin_num=20, min_radius=6.0, max_radius=28.0, n_estimators=80)
#ispline_algo = reference_implementations.InterpolatingSplineAlgorithm()
algos = [rf_algo, rfc_algo]
samples = automatic_test.load_samples('../data/training_data/samples.pickle')
samples = automatic_test.normalize_features(samples)
train_samples, test_samples = automatic_test.get_partitioned_samples(samples, 0.8)
automatic_test.train(algos, train_samples)
results = automatic_test.predict(algos, test_samples)
results_proba = automatic_test.predict_proba([rfc_algo], test_samples)
automatic_test.show_intersection_plot(results, test_samples, results_proba, which_samples="best-worst-case", orientation="curve-secant")
# automatic_test.show_graph_plot(results, test_samples, results_proba, which_samples="best-worst-case")
N_tests = 10
training_samples_steps = np.arange(10, N_training_samples+1, 10)
all_training_errors = []
all_cv_errors = []
all_training_samples_sizes = []
for test_i in range(N_tests):
print "====== Test number %d ======" % test_i
random.shuffle(samples)
cv_samples = samples[N_training_samples:]
for training_samples_size in training_samples_steps:
print "Training samples size:", training_samples_size
training_samples = samples[:training_samples_size]
automatic_test.train([rf_algo], training_samples)
rs_train = automatic_test.get_result_statistics(automatic_test.predict([rf_algo], training_samples))
rs_cv = automatic_test.get_result_statistics(automatic_test.predict([rf_algo], cv_samples))
all_training_samples_sizes.append(training_samples_size)
all_training_errors.append(rs_train[rf_algo]['mean_mse'])
all_cv_errors.append(rs_cv[rf_algo]['mean_mse'])
training_errors = []
cv_errors = []
training_samples_sizes = []
# Flatten the results
for training_samples_size in training_samples_steps:
indices = [i for i,x in enumerate(all_training_samples_sizes) if x == training_samples_size]
training_samples_sizes.append(training_samples_size)
