225 lines
9.4 KiB
Python
225 lines
9.4 KiB
Python
import numpy as np
|
|
import main
|
|
|
|
def test_step_01():
|
|
test = main.load_data('iris.csv')
|
|
assert test is not None
|
|
x,y = test
|
|
assert type(x) == np.ndarray
|
|
assert type(y) == np.ndarray
|
|
assert x.shape == (150,4)
|
|
assert y.shape == (150,)
|
|
assert x.dtype == np.float32 or x.dtype == np.float64
|
|
assert y.dtype == np.int32 or y.dtype == np.int64
|
|
|
|
def test_step_02():
|
|
x = np.array([
|
|
[59.76270078546494,93.0378732744839,70.55267521432877,],
|
|
[58.976636599379376,34.73095986778094,79.17882261333122,],
|
|
[37.5174422525385,128.35460015641596,142.73255210020585,],
|
|
[26.68830376515554,108.34500761653291,55.778983950580894,],
|
|
[63.60891221878646,135.1193276585322,-35.79278836042261,],
|
|
[-32.57414005969186,-45.956320511934855,116.5239691095876,],
|
|
[105.6313501899701,124.00242964936382,145.7236684465528,],
|
|
[109.83171284334472,42.29587245058637,106.10583525729109,],
|
|
[-26.345114826213354,77.98420426550476,-21.32934251819072,],
|
|
[138.93378340991677,54.36966435001433,32.932387998104716,],
|
|
])
|
|
x_ref = np.array([
|
|
[0.10485104208319312,0.3362373314630696,0.021964359573702855,],
|
|
[0.09002614318807045,-0.7645767522567821,0.166375923375568,],
|
|
[-0.31468684615052994,1.0030047535583193,1.2303379722621464,],
|
|
[-0.518920655600867,0.625230661423531,-0.22536411278355217,],
|
|
[0.17738926051062492,1.1307204372014275,-1.7583802123078445,],
|
|
[-1.636589865740423,-2.2879243182562656,0.7915763467485558,],
|
|
[0.9699179181914451,0.9208373001811287,1.2804126775864766,],
|
|
[1.0491353062386208,-0.6217538548213571,0.6171648821726928,],
|
|
[-1.5191126025141926,0.05202933813039103,-1.5162456027970177,],
|
|
[1.5979902997940605,-0.39380489662345974,-0.607842233830729,],
|
|
])
|
|
x_test = main.standardize(x)
|
|
assert x_test is not None
|
|
error = np.max(np.abs(x_ref - x_test))
|
|
assert x_test.shape == x_ref.shape
|
|
assert error < 1e-6
|
|
|
|
def test_step_03():
|
|
x = np.array([
|
|
[-1.105742368566682,1.2447972510784915,-0.10292263179401698,],
|
|
[0.5207869535408698,-1.2813536677837922,0.5399586812056,],
|
|
[0.7544895471006708,0.7187982077571432,1.838230783931978,],
|
|
[1.1276950515239543,-0.14201644595920881,-0.1790339466010727,],
|
|
[1.2123239446742684,-1.1427957147947074,0.7355527607135005,],
|
|
[1.0678403424190925,-0.6406810309665311,-1.405619490464019,],
|
|
[-0.8334425979838738,-0.12796234478939209,0.3511025041846954,],
|
|
[-0.17414992612560526,1.960866758149327,-1.5126361666589871,],
|
|
[-1.4037672060254274,-0.80477781558321,0.6811779598355046,],
|
|
[-1.1660337405572676,0.2151248028918804,-0.9458104543531817,],
|
|
])
|
|
C_ref = np.array([
|
|
[0.9999999999999998,-0.30871041083062684,0.10882239075522164,],
|
|
[-0.30871041083062684,1.0,-0.3175955960839056,],
|
|
[0.10882239075522164,-0.3175955960839056,1.0,],
|
|
])
|
|
C_test = main.compute_covariance_matrix(x)
|
|
assert C_test is not None
|
|
error = np.max(np.abs(C_ref - C_test))
|
|
assert C_test.shape == (3,3)
|
|
assert error < 1e-6
|
|
|
|
def test_step_04():
|
|
C = np.array([
|
|
[0.9999999999999998,-0.30871041083062684,0.10882239075522164,],
|
|
[-0.30871041083062684,1.0,-0.3175955960839056,],
|
|
[0.10882239075522164,-0.3175955960839056,1.0,],
|
|
])
|
|
eigval_ref = np.array([1.5006348637289717,0.8912274127999972,0.6081377234710301,])
|
|
eigvec_ref = np.array([
|
|
[-0.524750256529313,0.7183989075784977,0.4566619952026235,],
|
|
[0.6626030910538614,0.007925719552699393,0.7489287861308581,],
|
|
[-0.5344102469061465,-0.6955862222324367,0.4801723601396883,],
|
|
])
|
|
test = main.compute_eigenelements(C)
|
|
assert test is not None
|
|
eigval_test, eigvec_test = test
|
|
assert eigval_test.shape == (3,)
|
|
assert eigvec_test.shape == (3,3)
|
|
error = np.max(np.abs(eigval_ref - eigval_test))
|
|
assert error < 1e-6
|
|
error = np.max(np.abs(eigvec_ref - eigvec_test))
|
|
assert error < 1e-6
|
|
|
|
def test_step_05():
|
|
eigval = np.array([1.0, 3.0, 2.0])
|
|
eigvec = np.array([
|
|
[1.0, 3.0, 2.0],
|
|
[1.1, 3.1, 2.1],
|
|
[1.2, 3.2, 2.2],
|
|
])
|
|
eigval_ref = np.array([3.0, 2.0, 1.0])
|
|
eigvec_ref = np.array([
|
|
[3.0, 2.0, 1.0],
|
|
[3.1, 2.1, 1.1],
|
|
[3.2, 2.2, 1.2],
|
|
])
|
|
test = main.sort_eigenelements(eigval, eigvec)
|
|
assert test is not None
|
|
eigval_test, eigvec_test = test
|
|
assert eigval_test.shape == (3,)
|
|
assert eigvec_test.shape == (3,3)
|
|
error = np.max(np.abs(eigval_ref - eigval_test))
|
|
assert error < 1e-6
|
|
error = np.max(np.abs(eigvec_ref - eigvec_test))
|
|
assert error < 1e-6
|
|
|
|
def test_step_06():
|
|
eigval = np.array([0.9767610881903371,0.9764594650133958,0.8379449074988039,0.8209932298479351,0.7392635793983017,0.6563295894652734,0.604845519745046,0.4686512016477016,0.3687251706609641,0.29614019752214493,0.2828069625764096,0.1965823616800535,0.15896958364551972,0.1381829513486138,0.1201965612131689,0.11872771895424405,0.11037514116430513,0.09710127579306127,0.09609840789396307,0.039187792254320675,])
|
|
var_ref = np.array([0.12052317179600819,0.12048595432040264,0.10339455498701101,0.1013028767020668,0.09121820316104898,0.08098492540533334,0.07463227330373827,0.05782716978748539,0.045497233280881885,0.03654092728836284,0.03489573094977952,0.024256422614793195,0.019615358015076186,0.017050482237702617,0.014831130121310949,0.014649888740942673,0.013619258856022274,0.011981388166988682,0.011857643659193383,0.004835406605851003,])
|
|
var_accu_ref = np.array([0.12052317179600819,0.24100912611641084,0.34440368110342184,0.4457065578054886,0.5369247609665376,0.6179096863718709,0.6925419596756092,0.7503691294630945,0.7958663627439764,0.8324072900323393,0.8673030209821188,0.891559443596912,0.9111748016119882,0.9282252838496908,0.9430564139710018,0.9577063027119445,0.9713255615679667,0.9833069497349554,0.9951645933941488,0.9999999999999998,])
|
|
test = main.compute_var_ratio(eigval)
|
|
assert test is not None
|
|
ver_test, var_accu_test = test
|
|
assert ver_test.shape == eigval.shape
|
|
assert var_accu_test.shape == eigval.shape
|
|
error = np.max(np.abs(var_ref - ver_test))
|
|
assert error < 1e-6
|
|
error = np.max(np.abs(var_accu_ref - var_accu_test))
|
|
assert error < 1e-6
|
|
|
|
# def test_step_07():
|
|
# pass
|
|
|
|
def test_step_08():
|
|
x = np.array([
|
|
[-0.5692272135195057,-0.03821691441797431,-1.391429749664207,],
|
|
[0.8349483159428722,0.5897352401152248,-0.7587336993772015,],
|
|
[0.2004294637899314,-1.358613577394327,0.21764397273661457,],
|
|
[1.722938694045615,-0.43267593685524763,0.5052024345091827,],
|
|
[-1.2673436013591408,1.2069175815704556,-0.6832265505709595,],
|
|
[-1.0746396605693356,0.671812065430423,-1.529889431394962,],
|
|
[1.3466454025825358,-1.726489239590098,0.5379189375433167,],
|
|
[-0.7491139962207634,1.2846882183405577,1.4319721238227643,],
|
|
[-0.8288106401604025,0.6291253952950716,0.26824732034137383,],
|
|
[0.38417323546819404,-0.8262828324940856,1.4022946420540787,],
|
|
])
|
|
eigvec = np.array([
|
|
[-0.6432311709554334,0.725040765848229,],
|
|
[0.6234976274295615,0.6825615590553287,],
|
|
[-0.44441463668629316,-0.09179109955857746,],
|
|
])
|
|
proj_ref = np.array([
|
|
[0.9606882782960829,-0.3110774648390586,],
|
|
[0.16782610142261073,1.077547171895806,],
|
|
[-1.072738987857223,-0.8019956491849984,],
|
|
[-1.602539650115925,0.9074997412514691,],
|
|
[1.8713410365637415,-0.032366112869744654,],
|
|
[1.7900202119211768,-0.18017423862691137,],
|
|
[-2.1817252930269992,-0.2514385438029105,],
|
|
[0.6464641579366405,0.20229831190009098,],
|
|
[0.8061619944812769,-0.1961274071189318,],
|
|
[-1.385497849621383,-0.41416580860481034,],
|
|
])
|
|
proj_test = main.project_on_2pc(x, eigvec)
|
|
assert proj_test is not None
|
|
assert proj_test.shape == (len(x), 2)
|
|
error = np.max(np.abs(proj_ref - proj_test))
|
|
assert error < 1e-6
|
|
|
|
# def test_step_09():
|
|
# pass
|
|
|
|
# def test_step_10():
|
|
# pass
|
|
|
|
def test_step_11():
|
|
x = np.array([
|
|
[-2.0, -100.0],
|
|
[-2.0, +100.0],
|
|
[ 0.0, -100.0],
|
|
[ 0.0, +100.0],
|
|
[+2.0, -100.0],
|
|
[+2.0, +100.0],
|
|
])
|
|
y_ref = np.array([0,0, 1,1, 2,2], dtype=int)
|
|
y_test = main.classify(x)
|
|
assert y_test is not None
|
|
assert y_test.shape == (6,)
|
|
assert y_test.dtype == np.int32 or y_test.dtype == np.int64
|
|
# print(y_ref, y_test)
|
|
# assert all(y_ref[i] == y_test[i] for i in range(len(x)))
|
|
error = np.max(np.abs(y_ref - y_test))
|
|
assert error == 0
|
|
|
|
def test_step_12():
|
|
y = np.array([0,0,0,0, 1,1,1,1, 2,2,2,2])
|
|
y_pred = np.array([0,0,1,2, 1,1,1,1, 2,2,2,1])
|
|
conf_ref = np.array([
|
|
[2,1,1,],
|
|
[0,4,0,],
|
|
[0,1,3,],
|
|
], dtype=int)
|
|
conf_test = main.compute_confusion_matrix(y, y_pred)
|
|
assert conf_test is not None
|
|
assert conf_test.shape == (3,3)
|
|
assert conf_test.dtype == np.int32 or conf_test.dtype == np.int64
|
|
error = np.max(np.abs(conf_ref - conf_test))
|
|
assert error == 0
|
|
|
|
def test_step_13():
|
|
conf = np.array([
|
|
[2,1,1,],
|
|
[0,4,0,],
|
|
[0,1,3,],
|
|
], dtype=int)
|
|
prec_ref = 0.8055555555555555
|
|
rec_ref = 0.75
|
|
test = main.compute_metric(conf)
|
|
assert test is not None
|
|
prec_test, rec_test = test
|
|
error = np.max(np.abs(prec_ref - prec_test))
|
|
assert error < 1e-6
|
|
error = np.max(np.abs(rec_ref - rec_test))
|
|
assert error < 1e-6
|
|
|
|
# def test_step_14():
|
|
# pass
|