import numpy as np
from sklearn import preprocessing
from sklearn.decomposition import PCA
from evaluation.compare import compare, compare_selected, classifiers_all
from preprocessing.PreProcessor import PreProcessor
from src.filereader.FileReader import FileReader
FILE_PATH_1 = "../../resource/PostureEntry_DMP_Phil_Monday.csv"
FILE_PATH_2 = "../../resource/PostureEntry_DMP_Sergio_Monday.csv"
FILE_PATH_3 = "../../resource/PostureEntry_DMP_Ozan.csv"
N_VALIDATIONS = 30
samples_raw, labels_raw, label_names = FileReader.readAll(
[FILE_PATH_1, FILE_PATH_2, FILE_PATH_3])
scores = []
samples_filtered, labels_reduced = PreProcessor.magnitude_theta(
samples_raw, labels_raw, 5)
samples_features = preprocessing.scale(samples_filtered)
# samples_features = PCA().fit_transform(samples_features)
compare(samples_features, labels_reduced, classifiers_all, N_VALIDATIONS)
import numpy as np from sklearn import preprocessing from evaluation.compare import compare from preprocessing.PreProcessor import PreProcessor from src.filereader.FileReader import FileReader FILE_PATH = "../../resource/PostureEntry.csv" N_VALIDATIONS = 4 samples_raw, labels, label_names = FileReader.read(FILE_PATH) labels = labels[(labels == 4) | (labels == 5)] samples_raw = samples_raw[(labels == 4) | (labels == 5), :] # determine g when calibrating # determine upright angle # put threshold on angles samples_features = preprocessing.scale(samples_raw) compare(samples_features, labels, N_VALIDATIONS)
import numpy as np from sklearn import preprocessing from evaluation.compare import compare from preprocessing.PreProcessor import PreProcessor from src.filereader.FileReader import FileReader FILE_PATH = "../../resource/PostureEntry.csv" N_VALIDATIONS = 4 FILE_PATH_1 = "../../resource/PostureEntry_11_02_back_sit_stand_phil.csv" samples_raw, labels, _ = FileReader.read(FILE_PATH_1) samples_raw = samples_raw[:, 0:6] window_size = 50 window = PreProcessor.merge_window(samples_raw, window_size) # p2p = PreProcessor.peak2peak(samples_raw, window_size) labels_reduced = labels[0::window_size] labels_reduced = labels_reduced[:-1] samples_features = window samples_features = preprocessing.scale(samples_features) compare(samples_features, labels_reduced, N_VALIDATIONS)
from sklearn import preprocessing
from sklearn.decomposition import PCA
from sklearn.feature_selection import SelectKBest
from sklearn.model_selection import cross_val_score
from sklearn.neighbors import KNeighborsClassifier
from sklearn.pipeline import Pipeline
from evaluation.compare import compare, compare_selected, classifiers_all
from preprocessing.PreProcessor import PreProcessor
from src.filereader.FileReader import FileReader
FILE_PATH_1 = "../../resource/PostureEntry_DMP_Phil_Monday.csv"
FILE_PATH_2 = "../../resource/PostureEntry_DMP_Sergio_Monday.csv"
FILE_PATH_3 = "../../resource/PostureEntry_DMP_Ozan.csv"
paths = [FILE_PATH_1, FILE_PATH_2, FILE_PATH_3]
N_VALIDATIONS = 30
# samples_filtered, labels_reduced = PreProcessor.average(samples_raw, labels_raw, window_size)
scores_mean = 0.0
scores_std = 0.0
for path in paths:
samples_raw, labels_raw, label_names = FileReader.readAll([path])
samples_features = preprocessing.scale(samples_raw)
scores_clf = cross_val_score(KNeighborsClassifier(5, metric='manhattan', algorithm='kd_tree'), samples_features,
labels_raw.ravel(), cv=N_VALIDATIONS)
scores_mean += scores_clf.mean()
scores_std += scores_clf.std()
print("Mean:" + str(scores_mean/3))
print("Std:" + str(scores_std/3))
import numpy as np
from sklearn import preprocessing
from evaluation.compare import compare
from preprocessing.PreProcessor import PreProcessor
from src.filereader.FileReader import FileReader
FILE_PATH = "../../resource/PostureEntry.csv"
N_VALIDATIONS = 4
# FILE_PATH = "../../../../App/Generated-Data/PostureEntry_11_02_chest_sitting.csv"
FILE_PATH_1 = "../../resource/PostureEntry_11_02_back_sit_move_sergio.csv"
# FILE_PATH_2 = "../../resource/PostureEntry_11_02_back_phil_walk_outside.csv"
FILE_PATH_3 = "../../resource/PostureEntry_11_02_chest_sitting_phil.csv"
samples_raw_1, labels_1, _ = FileReader.read(FILE_PATH_1)
samples_raw_1 = samples_raw_1[(labels_1.ravel() == 0) |
(labels_1.ravel() == 1), :]
labels_1 = labels_1[(labels_1.ravel() == 0) | (labels_1.ravel() == 1)]
#samples_raw_2, labels_2, _ = FileReader.read(FILE_PATH_2)
samples_raw_3, labels_3, _ = FileReader.read(FILE_PATH_3)
samples_raw = np.vstack([samples_raw_1, samples_raw_3])
labels = np.vstack([labels_1, labels_3])
window_size = 20
window = PreProcessor.merge_window(samples_raw, window_size)
# p2p = PreProcessor.peak2peak(samples_raw, window_size)
labels_reduced = labels[0::window_size]
labels_reduced = labels_reduced[:-1]
import numpy as np
from sklearn import preprocessing
from sklearn.decomposition import PCA
from evaluation.compare import compare, compare_selected
from preprocessing.PreProcessor import PreProcessor
from src.filereader.FileReader import FileReader
# FILE_PATH = "../../../../App/Generated-Data/PostureEntry_11_02_chest_sitting.csv"
FILE_PATH_1 = "../../resource/PostureEntry_DMP_Phil_Monday.csv"
FILE_PATH_2 = "../../resource/PostureEntry_DMP_Sergio_Monday.csv"
N_VALIDATIONS = 20
samples_raw, labels_raw, _ = FileReader.readAll([FILE_PATH_1])
window_size = 20
average, labels_reduced = PreProcessor.median(samples_raw, labels_raw,
window_size)
p2p = PreProcessor.peak2peak(samples_raw, window_size)
samples_features = np.zeros(
shape=(int(np.floor(samples_raw.shape[0] / window_size)), 9))
samples_features[:, 0:9] = average[:, 0:9]
# samples_features[:, 6] = p2p.ravel()
samples_features = preprocessing.scale(samples_features)
# samples_features = PCA().fit_transform(samples_features)
compare_selected(samples_features, labels_reduced, N_VALIDATIONS)