def load_yearpred(folder="../data/"):
data = np.genfromtxt(os.path.join(folder, 'YearPredictionMSD.txt'), delimiter=',')
test_X = data[463715:, 1:]
test_y = data[463715:, 0]
validation_X = data[410000:463715, 1:]
validation_y = data[410000:463715, 0]
train = munge.shuffle_data(data[0:410000, :])
train_X = train[:, 1:]
train_y = train[:, 0]
return train_X, train_y, validation_X, validation_y, test_X, test_y
def load_seismic_station(siteid, phaseid=0, folder="../data/"):
reader = csv.reader(open(os.path.join(folder, 'tt_data.csv'), 'rb'), delimiter=',')
data = np.array([[float(col) for col in row] for row in reader if int(row[0])==int(siteid)] and int(row[1])==int(phaseid))
data = munge.shuffle_data(data)
train_n = .8 * data.shape[0]
validate_n = .1 * data.shape[0]
# cols: [siteid, phaseid, evlon, evlat, evdepth, sitelon, sitelat, siteheight, tt, ttres]
X = data[0:train_n, [2,3,4,5,6,7] ]
y = data[0:train_n, 9]
v_cutoff = train_n+validate_n
validation_X = data[train_n:v_cutoff, [2,3,4,5,6,7] ]
validation_y = data[train_n:v_cutoff, 9]
test_X = data[v_cutoff:, [2,3,4,5,6,7] ]
test_y = data[v_cutoff:, 9]
return X, y, validation_X, validation_y, test_X, test_y
