Perceptron

import pandas spam_all = pandas.read_table("spambase.txt",sep=",",header=0,decimal=".")

from sklearn import metrics

spam_all.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 4601 entries, 0 to 4600
Data columns (total 58 columns):
 #   Column                      Non-Null Count  Dtype  
---  ------                      --------------  -----  
 0   word_freq_make              4601 non-null   float64
 1   word_freq_address           4601 non-null   float64
 2   word_freq_all               4601 non-null   float64
 3   word_freq_3d                4601 non-null   float64
 4   word_freq_our               4601 non-null   float64
 5   word_freq_over              4601 non-null   float64
 6   word_freq_remove            4601 non-null   float64
 7   word_freq_internet          4601 non-null   float64
 8   word_freq_order             4601 non-null   float64
 9   word_freq_mail              4601 non-null   float64
 10  word_freq_receive           4601 non-null   float64
 11  word_freq_will              4601 non-null   float64
 12  word_freq_people            4601 non-null   float64
 13  word_freq_report            4601 non-null   float64
 14  word_freq_addresses         4601 non-null   float64
 15  word_freq_free              4601 non-null   float64
 16  word_freq_business          4601 non-null   float64
 17  word_freq_email             4601 non-null   float64
 18  word_freq_you               4601 non-null   float64
 19  word_freq_credit            4601 non-null   float64
 20  word_freq_your              4601 non-null   float64
 21  word_freq_font              4601 non-null   float64
 22  word_freq_000               4601 non-null   float64
 23  word_freq_money             4601 non-null   float64
 24  word_freq_hp                4601 non-null   float64
 25  word_freq_hpl               4601 non-null   float64
 26  word_freq_george            4601 non-null   float64
 27  word_freq_650               4601 non-null   float64
 28  word_freq_lab               4601 non-null   float64
 29  word_freq_labs              4601 non-null   float64
 30  word_freq_telnet            4601 non-null   float64
 31  word_freq_857               4601 non-null   float64
 32  word_freq_data              4601 non-null   float64
 33  word_freq_415               4601 non-null   float64
 34  word_freq_85                4601 non-null   float64
 35  word_freq_technology        4601 non-null   float64
 36  word_freq_1999              4601 non-null   float64
 37  word_freq_parts             4601 non-null   float64
 38  word_freq_pm                4601 non-null   float64
 39  word_freq_direct            4601 non-null   float64
 40  word_freq_cs                4601 non-null   float64
 41  word_freq_meeting           4601 non-null   float64
 42  word_freq_original          4601 non-null   float64
 43  word_freq_project           4601 non-null   float64
 44  word_freq_re                4601 non-null   float64
 45  word_freq_edu               4601 non-null   float64
 46  word_freq_table             4601 non-null   float64
 47  word_freq_conference        4601 non-null   float64
 48  char_freq_;                 4601 non-null   float64
 49  char_freq_(                 4601 non-null   float64
 50  char_freq_[                 4601 non-null   float64
 51  char_freq_!                 4601 non-null   float64
 52  char_freq_$                 4601 non-null   float64
 53  char_freq_#                 4601 non-null   float64
 54  capital_run_length_average  4601 non-null   float64
 55  capital_run_length_longest  4601 non-null   int64  
 56  capital_run_length_total    4601 non-null   int64  
 57  spam                        4601 non-null   int64  
dtypes: float64(55), int64(3)
memory usage: 2.0 MB

from sklearn.model_selection import train_test_split X_train, X_test, Y_train, Y_test = train_test_split(spam_all.iloc[:,0:57],spam_all.loc[:,"spam"], test_size=0.2175)

from sklearn.preprocessing import StandardScaler scaler = StandardScaler(with_mean=True,with_std=True) scaler.fit(X_train) Z_train = scaler.transform(X_train) print(Z_train.mean(axis=0))

[ 7.10542736e-17 -1.57898386e-17 -4.90965293e-17 -7.40148683e-18
 -6.71068139e-17  7.69754630e-17  3.99680289e-17 -6.01987596e-17
 -2.81256500e-17  2.46716228e-18 -4.04614613e-17  3.35534070e-17
 -2.22044605e-17  2.96059473e-18  3.65140017e-17  0.00000000e+00
  4.73695157e-17  1.97372982e-17 -2.31913254e-17 -1.48029737e-17
  1.06581410e-16  3.45402719e-17 -1.38161088e-17  4.44089210e-17
  1.97372982e-18  6.90805438e-18 -1.18423789e-17  1.77635684e-17
  3.25665421e-17 -2.96059473e-18  5.92118946e-18 -9.37521665e-18
 -2.91125149e-17 -9.86864911e-18 -1.97372982e-18 -8.38835174e-18
  2.17110280e-17  4.93432455e-18  9.37521665e-18  4.44089210e-17
  1.18423789e-17  6.90805438e-18 -3.84877315e-17 -9.86864911e-18
 -6.80936788e-17 -4.24351912e-17 -2.07241631e-17 -2.17110280e-17
 -9.86864911e-18  1.57898386e-17 -1.87504333e-17  3.84877315e-17
 -2.56584877e-17 -5.92118946e-18  1.97372982e-18  9.86864911e-18
 -2.86190824e-17]

Z_test = scaler.transform(X_test) print(Z_test.mean(axis=0))

[-0.0378188   0.00358165  0.06467463 -0.03443819  0.00012075  0.01759126
 -0.00263605  0.02212304 -0.00288727 -0.02173335 -0.01619425 -0.02070032
  0.00399926 -0.01111788  0.00740293 -0.03752886 -0.04083832 -0.01174155
  0.00854839  0.00209437 -0.05678322  0.01281356  0.06288376 -0.04528819
 -0.00078119 -0.00640172  0.00746968 -0.04533094  0.045163   -0.00939345
  0.0515831   0.04010162  0.03448559  0.03836718  0.0050869  -0.00329967
 -0.05297989  0.02497115 -0.05907397  0.05117547 -0.00475573  0.05420181
  0.0312495   0.06407833  0.01715653  0.05280285 -0.0275301  -0.0271555
 -0.0098211   0.00911014  0.01566219  0.02671501  0.02646856 -0.04073214
  0.02068994  0.02206846  0.04057584]

def eval_model(y_true,y_pred,poslabel="yes"): #matrice de confusion mc = metrics.confusion_matrix(y_true,y_pred) print("Confusion matrix :") print(mc) #taux d'erreur err = 1 - metrics.accuracy_score(y_true,y_pred) print("Err-rate = ",err) #F1-score f1 = metrics.f1_score(y_true,y_pred) print("F1-Score = ",f1)

from sklearn.linear_model import Perceptron ps_sklearn = Perceptron(random_state=100, max_iter = 1500, tol = None) ps_sklearn.fit(Z_train,Y_train) preds_sklearn = ps_sklearn.predict(Z_test) eval_model(Y_test,preds_sklearn) print(ps_sklearn.intercept_) coefs = ps_sklearn.coef_[0,:] print(coefs)

Confusion matrix :
[[536  47]
 [ 80 338]]
Err-rate =  0.1268731268731269
F1-Score =  0.8418430884184308
[-127.]
[-1.13796959e+01 -4.25097686e+00  6.59798995e+00  3.81698345e+01
 -7.05297569e-01 -2.08216101e+00  1.38055033e+01  2.24235261e+00
  5.25409317e+00 -1.66195405e-03 -6.67012069e+00 -4.34682826e-01
 -1.65223151e+00  7.56442062e-01  2.48602899e+00  9.99889858e+00
  9.99682046e+00 -4.49292163e+00 -1.10954474e+00  1.03085781e+01
 -3.28016234e+00  3.45401716e+00  8.09630980e+00  1.44315768e+00
 -2.94452445e+01 -7.51547805e+00 -4.10397932e+02  3.73780322e+00
 -1.35463792e+01 -5.82628853e+00 -1.01578215e+01  1.19890646e+01
  7.86598548e-01 -2.31751454e+00 -1.28862027e+01  5.92507341e+00
  8.42491898e+00 -1.84117464e+00 -4.16355123e+00 -1.95568343e+00
 -1.65036794e+02 -1.47217462e+01 -1.34100152e+00 -3.20492105e+00
 -7.64457453e+00 -7.25034234e+00 -4.79431435e+00 -1.26801157e+01
 -9.54395451e+00 -9.91854324e+00 -1.54084209e+00  5.17460391e+00
  1.29701951e+01  9.43505542e+00 -1.37494604e-02  1.91517326e+01
  2.72050143e+00]

from sklearn.neural_network import MLPClassifier pmc_sklearn = MLPClassifier(hidden_layer_sizes=(2),activation='logistic',max_iter=1500) pmc_sklearn.fit(Z_train,Y_train) preds_sklearn = pmc_sklearn.predict(Z_test) eval_model(Y_test,preds_sklearn) print(pmc_sklearn.coefs_)

Confusion matrix :
[[552  31]
 [ 43 375]]
Err-rate =  0.07392607392607398
F1-Score =  0.9101941747572816
[array([[-0.06243585, -0.06972204],
       [ 0.13250031,  0.87436484],
       [ 0.21657114,  0.10636347],
       [-1.36903989, -1.07360363],
       [-1.19238625,  0.0240723 ],
       [-0.3361549 ,  0.02593657],
       [-1.62835831, -1.67187932],
       [-0.65107769,  0.04250276],
       [-0.09706641, -0.10477043],
       [ 0.07185751, -0.35673139],
       [ 0.2916899 , -1.4292945 ],
       [ 0.25042177,  0.44882076],
       [ 0.32627815,  0.15686032],
       [-0.03986608,  0.38748051],
       [ 0.04443746, -1.55455247],
       [-1.13080742, -0.74412041],
       [-0.87798125, -0.09927781],
       [-0.0551511 ,  0.16728481],
       [ 0.48342836, -0.34262385],
       [-0.77003348, -1.17587602],
       [ 0.13943419, -1.05115761],
       [-0.32932729, -0.12943647],
       [-1.21986323, -1.97027382],
       [-1.08311562, -0.6709295 ],
       [ 2.84898644,  3.22994563],
       [ 1.07815561,  1.04307161],
       [ 2.79581038,  3.53266315],
       [-0.9735323 ,  0.47792815],
       [ 0.62755065,  0.89005886],
       [-0.2101394 ,  0.68451678],
       [ 1.68139129,  0.81357101],
       [ 0.54515367,  0.73424091],
       [ 0.75527872, -0.06756066],
       [ 0.41776336,  1.06760761],
       [ 1.33988984,  1.18188638],
       [-0.40609229, -0.59174301],
       [ 2.09832975, -0.31277529],
       [ 0.16210801,  0.36809088],
       [ 0.26527651,  0.19144769],
       [ 0.18367503,  0.20975213],
       [ 1.81743679,  1.60414177],
       [ 1.07415694,  1.60767341],
       [-0.12100138,  0.71066534],
       [ 0.88311085,  1.65076174],
       [ 0.87540491,  0.92428849],
       [ 0.12507529,  3.58442148],
       [ 0.36745558,  0.44790636],
       [ 1.09135617,  0.39265451],
       [ 0.39152106,  0.17010817],
       [ 0.04839576,  0.08776854],
       [ 0.25712437,  0.57072881],
       [-0.84029151, -2.47133243],
       [-1.65533065, -2.84921178],
       [-0.33349119, -0.19547485],
       [-1.28334094, -3.38925692],
       [-1.86988566, -1.06203154],
       [-1.27045417,  0.01986002]]), array([[-4.15219865],
       [-4.07585643]])]

print(pmc_sklearn.intercepts_)

[array([1.07143955, 1.17587178]), array([3.26533536])]

import numpy Z1 = numpy.zeros(Z_test.shape[0]) for i in range(Z1.shape[0]): Z1[i] = numpy.sum(Z_test[i,:]*pmc_sklearn.coefs_[0][:,0]) + pmc_sklearn.intercepts_[0][0] Z2 = numpy.zeros(Z_test.shape[0]) for i in range(Z2.shape[0]): Z2[i] = numpy.sum(Z_test[i,:]*pmc_sklearn.coefs_[0][:,1]) + pmc_sklearn.intercepts_[0][1] Z1 = 1.0/(1.0+numpy.exp(-Z1)) Z2 = 1.0/(1.0+numpy.exp(-Z2)) import matplotlib.pyplot as plt plt.scatter(Z1,Z2)

horiz = numpy.linspace(0,1) vertic = -pmc_sklearn.coefs_[1][0]/pmc_sklearn.coefs_[1][1] * horiz - pmc_sklearn.intercepts_[1]/pmc_sklearn.coefs_[1][1] plt.scatter(Z1,Z2, c=pandas.Series(['red', 'blue'])[Y_test], s=5) plt.plot(horiz, vertic, c="black", linewidth="3", linestyle="--") plt.show()

import keras print(keras.__version__)

2.4.3

from keras.models import Sequential from keras.layers import Dense #instanciation de la structure pmc_keras = Sequential() #architeture du réseau #entree vers couche cachée pmc_keras.add(Dense(units=2,input_dim=57,activation="sigmoid")) #couche cachée vers sortie pmc_keras.add(Dense(units=1,activation="sigmoid"))

pmc_keras.compile(loss="binary_crossentropy",optimizer="adam",metrics=["accuracy"])

pmc_keras.fit(Z_train,Y_train.astype("float"),epochs=150,batch_size=15)

Epoch 1/150
240/240 [==============================] - 0s 850us/step - loss: 0.7042 - accuracy: 0.5243
Epoch 2/150
240/240 [==============================] - 0s 1ms/step - loss: 0.5477 - accuracy: 0.7971
Epoch 3/150
240/240 [==============================] - 0s 1ms/step - loss: 0.4713 - accuracy: 0.8570
Epoch 4/150
240/240 [==============================] - 0s 889us/step - loss: 0.4112 - accuracy: 0.8963
Epoch 5/150
240/240 [==============================] - 0s 856us/step - loss: 0.3729 - accuracy: 0.9010
Epoch 6/150
240/240 [==============================] - 0s 1ms/step - loss: 0.3415 - accuracy: 0.9101
Epoch 7/150
240/240 [==============================] - 0s 931us/step - loss: 0.3118 - accuracy: 0.9159
Epoch 8/150
240/240 [==============================] - 0s 1ms/step - loss: 0.3051 - accuracy: 0.9155
Epoch 9/150
240/240 [==============================] - 0s 892us/step - loss: 0.2894 - accuracy: 0.9127
Epoch 10/150
240/240 [==============================] - 0s 867us/step - loss: 0.2744 - accuracy: 0.9209
Epoch 11/150
240/240 [==============================] - 0s 954us/step - loss: 0.2497 - accuracy: 0.9310
Epoch 12/150
240/240 [==============================] - 0s 849us/step - loss: 0.2360 - accuracy: 0.9323
Epoch 13/150
240/240 [==============================] - 0s 1ms/step - loss: 0.2374 - accuracy: 0.9297
Epoch 14/150
240/240 [==============================] - 0s 885us/step - loss: 0.2251 - accuracy: 0.9326
Epoch 15/150
240/240 [==============================] - 0s 929us/step - loss: 0.2320 - accuracy: 0.9232
Epoch 16/150
240/240 [==============================] - 0s 880us/step - loss: 0.2288 - accuracy: 0.9265
Epoch 17/150
240/240 [==============================] - 0s 878us/step - loss: 0.2213 - accuracy: 0.9326
Epoch 18/150
240/240 [==============================] - 0s 1ms/step - loss: 0.2128 - accuracy: 0.9331
Epoch 19/150
240/240 [==============================] - 0s 934us/step - loss: 0.1965 - accuracy: 0.9401
Epoch 20/150
240/240 [==============================] - 0s 872us/step - loss: 0.2020 - accuracy: 0.9367
Epoch 21/150
240/240 [==============================] - 0s 858us/step - loss: 0.2085 - accuracy: 0.9345
Epoch 22/150
240/240 [==============================] - 0s 878us/step - loss: 0.2093 - accuracy: 0.9318
Epoch 23/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1906 - accuracy: 0.9404
Epoch 24/150
240/240 [==============================] - 0s 896us/step - loss: 0.2030 - accuracy: 0.9368
Epoch 25/150
240/240 [==============================] - 0s 899us/step - loss: 0.2034 - accuracy: 0.9309
Epoch 26/150
240/240 [==============================] - 0s 883us/step - loss: 0.2018 - accuracy: 0.9330
Epoch 27/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1964 - accuracy: 0.9343
Epoch 28/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1950 - accuracy: 0.9359
Epoch 29/150
240/240 [==============================] - 0s 966us/step - loss: 0.1982 - accuracy: 0.9335
Epoch 30/150
240/240 [==============================] - 0s 891us/step - loss: 0.2025 - accuracy: 0.9326
Epoch 31/150
240/240 [==============================] - 0s 869us/step - loss: 0.1915 - accuracy: 0.9376
Epoch 32/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1939 - accuracy: 0.9373
Epoch 33/150
240/240 [==============================] - 0s 963us/step - loss: 0.2066 - accuracy: 0.9323
Epoch 34/150
240/240 [==============================] - 0s 876us/step - loss: 0.1971 - accuracy: 0.9340
Epoch 35/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1862 - accuracy: 0.9383
Epoch 36/150
240/240 [==============================] - 0s 895us/step - loss: 0.1943 - accuracy: 0.9335
Epoch 37/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1979 - accuracy: 0.9342
Epoch 38/150
240/240 [==============================] - 0s 869us/step - loss: 0.1933 - accuracy: 0.9345
Epoch 39/150
240/240 [==============================] - 0s 896us/step - loss: 0.1906 - accuracy: 0.9377
Epoch 40/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1953 - accuracy: 0.9333
Epoch 41/150
240/240 [==============================] - 0s 878us/step - loss: 0.1896 - accuracy: 0.9336
Epoch 42/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1845 - accuracy: 0.9372
Epoch 43/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1820 - accuracy: 0.9349
Epoch 44/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1789 - accuracy: 0.9408
Epoch 45/150
240/240 [==============================] - 0s 881us/step - loss: 0.1760 - accuracy: 0.9446
Epoch 46/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1799 - accuracy: 0.9389
Epoch 47/150
240/240 [==============================] - 0s 920us/step - loss: 0.1860 - accuracy: 0.9387
Epoch 48/150
240/240 [==============================] - 0s 937us/step - loss: 0.1752 - accuracy: 0.9407
Epoch 49/150
240/240 [==============================] - 0s 857us/step - loss: 0.1661 - accuracy: 0.9443
Epoch 50/150
240/240 [==============================] - 0s 882us/step - loss: 0.1619 - accuracy: 0.9472
Epoch 51/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1790 - accuracy: 0.9409
Epoch 52/150
240/240 [==============================] - 0s 887us/step - loss: 0.1832 - accuracy: 0.9358
Epoch 53/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1711 - accuracy: 0.9413
Epoch 54/150
240/240 [==============================] - 0s 912us/step - loss: 0.1720 - accuracy: 0.9405
Epoch 55/150
240/240 [==============================] - 0s 937us/step - loss: 0.1783 - accuracy: 0.9407
Epoch 56/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1732 - accuracy: 0.9423
Epoch 57/150
240/240 [==============================] - 0s 920us/step - loss: 0.1815 - accuracy: 0.9406
Epoch 58/150
240/240 [==============================] - 0s 887us/step - loss: 0.1692 - accuracy: 0.9459
Epoch 59/150
240/240 [==============================] - 0s 878us/step - loss: 0.1654 - accuracy: 0.9466
Epoch 60/150
240/240 [==============================] - 0s 904us/step - loss: 0.1819 - accuracy: 0.9407
Epoch 61/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1901 - accuracy: 0.9362
Epoch 62/150
240/240 [==============================] - 0s 884us/step - loss: 0.1881 - accuracy: 0.9329
Epoch 63/150
240/240 [==============================] - 0s 874us/step - loss: 0.1823 - accuracy: 0.9358
Epoch 64/150
240/240 [==============================] - 0s 922us/step - loss: 0.1843 - accuracy: 0.9344
Epoch 65/150
240/240 [==============================] - 0s 907us/step - loss: 0.1579 - accuracy: 0.9447
Epoch 66/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1631 - accuracy: 0.9454
Epoch 67/150
240/240 [==============================] - 0s 904us/step - loss: 0.1635 - accuracy: 0.9463
Epoch 68/150
240/240 [==============================] - 0s 889us/step - loss: 0.1778 - accuracy: 0.9400
Epoch 69/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1643 - accuracy: 0.9459
Epoch 70/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1791 - accuracy: 0.9367
Epoch 71/150
240/240 [==============================] - 0s 928us/step - loss: 0.1584 - accuracy: 0.9460
Epoch 72/150
240/240 [==============================] - 0s 904us/step - loss: 0.1624 - accuracy: 0.9470
Epoch 73/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1723 - accuracy: 0.9444
Epoch 74/150
240/240 [==============================] - 0s 951us/step - loss: 0.1832 - accuracy: 0.9392
Epoch 75/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1644 - accuracy: 0.9477
Epoch 76/150
240/240 [==============================] - 0s 886us/step - loss: 0.1730 - accuracy: 0.9416
Epoch 77/150
240/240 [==============================] - 0s 892us/step - loss: 0.1831 - accuracy: 0.9394
Epoch 78/150
240/240 [==============================] - 0s 950us/step - loss: 0.1796 - accuracy: 0.9404
Epoch 79/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1686 - accuracy: 0.9447
Epoch 80/150
240/240 [==============================] - 0s 2ms/step - loss: 0.1744 - accuracy: 0.9428
Epoch 81/150
240/240 [==============================] - 0s 867us/step - loss: 0.1702 - accuracy: 0.9401
Epoch 82/150
240/240 [==============================] - 0s 937us/step - loss: 0.1618 - accuracy: 0.9456
Epoch 83/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1523 - accuracy: 0.9482
Epoch 84/150
240/240 [==============================] - 0s 2ms/step - loss: 0.1596 - accuracy: 0.9502
Epoch 85/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1398 - accuracy: 0.9547
Epoch 86/150
240/240 [==============================] - 0s 910us/step - loss: 0.1803 - accuracy: 0.9383
Epoch 87/150
240/240 [==============================] - 0s 879us/step - loss: 0.1738 - accuracy: 0.9432
Epoch 88/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1784 - accuracy: 0.9405
Epoch 89/150
240/240 [==============================] - 0s 964us/step - loss: 0.1651 - accuracy: 0.9449
Epoch 90/150
240/240 [==============================] - 0s 880us/step - loss: 0.1813 - accuracy: 0.9389
Epoch 91/150
240/240 [==============================] - 0s 873us/step - loss: 0.1581 - accuracy: 0.9512
Epoch 92/150
240/240 [==============================] - 0s 883us/step - loss: 0.1510 - accuracy: 0.9503
Epoch 93/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1733 - accuracy: 0.9403
Epoch 94/150
240/240 [==============================] - 0s 919us/step - loss: 0.1701 - accuracy: 0.9416
Epoch 95/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1494 - accuracy: 0.9513
Epoch 96/150
240/240 [==============================] - 0s 980us/step - loss: 0.1553 - accuracy: 0.9489
Epoch 97/150
240/240 [==============================] - 0s 901us/step - loss: 0.1665 - accuracy: 0.9408
Epoch 98/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1628 - accuracy: 0.9435
Epoch 99/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1711 - accuracy: 0.9428
Epoch 100/150
240/240 [==============================] - 0s 886us/step - loss: 0.1646 - accuracy: 0.9477
Epoch 101/150
240/240 [==============================] - 0s 880us/step - loss: 0.1584 - accuracy: 0.9482
Epoch 102/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1625 - accuracy: 0.9455
Epoch 103/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1548 - accuracy: 0.9479
Epoch 104/150
240/240 [==============================] - 0s 963us/step - loss: 0.1639 - accuracy: 0.9412
Epoch 105/150
240/240 [==============================] - 0s 933us/step - loss: 0.1448 - accuracy: 0.9536
Epoch 106/150
240/240 [==============================] - 0s 932us/step - loss: 0.1624 - accuracy: 0.9469
Epoch 107/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1482 - accuracy: 0.9488
Epoch 108/150
240/240 [==============================] - 0s 884us/step - loss: 0.1582 - accuracy: 0.9467
Epoch 109/150
240/240 [==============================] - 0s 878us/step - loss: 0.1650 - accuracy: 0.9476
Epoch 110/150
240/240 [==============================] - 0s 987us/step - loss: 0.1530 - accuracy: 0.9510
Epoch 111/150
240/240 [==============================] - 0s 930us/step - loss: 0.1627 - accuracy: 0.9416
Epoch 112/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1590 - accuracy: 0.9478
Epoch 113/150
240/240 [==============================] - 0s 955us/step - loss: 0.1499 - accuracy: 0.9484
Epoch 114/150
240/240 [==============================] - 0s 924us/step - loss: 0.1687 - accuracy: 0.9421
Epoch 115/150
240/240 [==============================] - 0s 980us/step - loss: 0.1625 - accuracy: 0.9435
Epoch 116/150
240/240 [==============================] - 0s 896us/step - loss: 0.1589 - accuracy: 0.9434
Epoch 117/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1519 - accuracy: 0.9487
Epoch 118/150
240/240 [==============================] - 0s 857us/step - loss: 0.1460 - accuracy: 0.9503
Epoch 119/150
240/240 [==============================] - 0s 938us/step - loss: 0.1438 - accuracy: 0.9533
Epoch 120/150
240/240 [==============================] - 0s 901us/step - loss: 0.1466 - accuracy: 0.9509
Epoch 121/150
240/240 [==============================] - 0s 962us/step - loss: 0.1636 - accuracy: 0.9452
Epoch 122/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1479 - accuracy: 0.9542
Epoch 123/150
240/240 [==============================] - 0s 895us/step - loss: 0.1607 - accuracy: 0.9464
Epoch 124/150
240/240 [==============================] - 0s 882us/step - loss: 0.1587 - accuracy: 0.9448
Epoch 125/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1461 - accuracy: 0.9532
Epoch 126/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1596 - accuracy: 0.9482
Epoch 127/150
240/240 [==============================] - 0s 910us/step - loss: 0.1526 - accuracy: 0.9505
Epoch 128/150
240/240 [==============================] - 0s 935us/step - loss: 0.1470 - accuracy: 0.9515
Epoch 129/150
240/240 [==============================] - 0s 990us/step - loss: 0.1539 - accuracy: 0.9512
Epoch 130/150
240/240 [==============================] - 0s 936us/step - loss: 0.1437 - accuracy: 0.9541
Epoch 131/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1611 - accuracy: 0.9472
Epoch 132/150
240/240 [==============================] - 0s 876us/step - loss: 0.1657 - accuracy: 0.9460
Epoch 133/150
240/240 [==============================] - 0s 871us/step - loss: 0.1533 - accuracy: 0.9491
Epoch 134/150
240/240 [==============================] - 0s 880us/step - loss: 0.1637 - accuracy: 0.9439
Epoch 135/150
240/240 [==============================] - 0s 917us/step - loss: 0.1500 - accuracy: 0.9498
Epoch 136/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1578 - accuracy: 0.9464
Epoch 137/150
240/240 [==============================] - 0s 899us/step - loss: 0.1505 - accuracy: 0.9531
Epoch 138/150
240/240 [==============================] - 0s 921us/step - loss: 0.1580 - accuracy: 0.9486
Epoch 139/150
240/240 [==============================] - 0s 885us/step - loss: 0.1636 - accuracy: 0.9452
Epoch 140/150
240/240 [==============================] - 0s 922us/step - loss: 0.1508 - accuracy: 0.9497
Epoch 141/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1530 - accuracy: 0.9508
Epoch 142/150
240/240 [==============================] - 0s 887us/step - loss: 0.1566 - accuracy: 0.9505
Epoch 143/150
240/240 [==============================] - 0s 873us/step - loss: 0.1547 - accuracy: 0.9506
Epoch 144/150
240/240 [==============================] - 0s 923us/step - loss: 0.1568 - accuracy: 0.9527
Epoch 145/150
240/240 [==============================] - 0s 891us/step - loss: 0.1627 - accuracy: 0.9458
Epoch 146/150
240/240 [==============================] - 0s 1ms/step - loss: 0.1559 - accuracy: 0.9487
Epoch 147/150
240/240 [==============================] - 0s 868us/step - loss: 0.1498 - accuracy: 0.9490
Epoch 148/150
240/240 [==============================] - 0s 913us/step - loss: 0.1537 - accuracy: 0.9494
Epoch 149/150
240/240 [==============================] - 0s 906us/step - loss: 0.1645 - accuracy: 0.9461
Epoch 150/150
240/240 [==============================] - 0s 985us/step - loss: 0.1597 - accuracy: 0.9468

proba_predm_keras = pmc_keras.predict(Z_test) #dimension print(proba_predm_keras.shape) #(1000,1) #10 premières valeurs print(proba_predm_keras[:10,:]) predm_keras = numpy.repeat(1,Z_test.shape[0]) predm_keras[proba_predm_keras[:,0] < 0.5] = 0 eval_model(Y_test,predm_keras)

(1001, 1)
[[0.771462  ]
 [0.00215307]
 [0.9225127 ]
 [0.00215563]
 [0.70727444]
 [0.00215653]
 [0.9694468 ]
 [0.23150545]
 [0.00215301]
 [0.01367447]]
Confusion matrix :
[[556  27]
 [ 41 377]]
Err-rate =  0.06793206793206796
F1-Score =  0.9172749391727495

print(pmc_keras.get_weights())

[array([[-0.21281983, -0.10157181],
       [ 0.08139528, -0.74373454],
       [-0.05763601, -0.22512633],
       [-1.2909224 ,  1.3671893 ],
       [ 0.04205126,  0.8585134 ],
       [-0.36352643, -0.06510309],
       [-1.1543103 ,  1.2757866 ],
       [-0.4628375 , -0.11004866],
       [-0.37497628, -0.13042518],
       [-0.8186402 , -0.25072148],
       [-1.4916872 , -0.37151438],
       [ 0.49916327, -0.04988626],
       [ 0.1420159 , -0.20927243],
       [ 0.3338182 ,  0.22187263],
       [-1.2657392 ,  0.05437269],
       [-0.9314528 ,  0.399455  ],
       [-0.39114612,  0.34221852],
       [-0.26668847, -0.24921793],
       [ 0.07881838,  0.02725873],
       [-0.9276161 ,  0.86684674],
       [-0.83217156, -0.1329859 ],
       [-0.24029939,  0.03876249],
       [-1.203158  ,  1.4351958 ],
       [-0.0646081 ,  1.5157334 ],
       [ 2.5468788 , -3.486414  ],
       [ 0.16601779, -2.224186  ],
       [ 3.3256993 , -3.3297658 ],
       [ 0.6505888 ,  1.0545496 ],
       [ 0.670919  , -0.58734   ],
       [-0.11943289,  0.0223447 ],
       [ 0.2683273 , -0.93258554],
       [ 0.2979921 , -0.6718159 ],
       [ 0.04215759, -0.11906405],
       [ 0.9371995 , -0.03606919],
       [ 0.6748611 , -1.4049692 ],
       [-0.40292063,  0.43167916],
       [ 1.0439144 ,  0.20460351],
       [ 0.51698035, -0.0682429 ],
       [ 0.11260268, -0.26121467],
       [ 0.23744704, -0.37326622],
       [ 1.057007  , -2.2303345 ],
       [ 1.1900942 , -1.2668478 ],
       [ 0.36424193, -0.04130829],
       [ 1.1576945 , -0.85325146],
       [ 0.47151244, -0.9230963 ],
       [ 3.093585  , -0.18473561],
       [ 0.29042947, -0.2738163 ],
       [ 0.28310764, -0.40340564],
       [ 0.23464772, -0.12674215],
       [ 0.8780015 ,  0.4846296 ],
       [ 0.02900665, -0.3931549 ],
       [-1.7615016 ,  0.39485896],
       [-2.1185894 ,  2.3174436 ],
       [-0.19490592,  0.12009343],
       [-0.52110606,  4.430523  ],
       [-0.35501477,  2.9799716 ],
       [-0.0560343 ,  0.17795672]], dtype=float32), array([ 0.8921996, -0.6820642], dtype=float32), array([[-4.9392695],
       [ 4.8738103]], dtype=float32), array([-1.1994672], dtype=float32)]