[DeepLearning] 이진분류 (유방암 데이터 분류 실습)

 

 

목표

• 환자 데이터를 바탕으로 유방암인지 아닌지를 구분해보기
• 딥러닝으로 이진분류 실습을 진행하기

 

In [1]:

# 라이브러리 불러오기

import numpy as np

import pandas as pd

import matplotlib.pyplot as plt

from sklearn.datasets import load_breast_cancer # 사이킷런에 내장되어 있는 유방암 데이터 가져오기

In [2]:

# 데이터 불러오기

breast_data = load_breast_cancer()

print(breast_data)

# 머신러닝 데이터구조: 번치객체(딕셔너리 형태)

 

{'data': array([[1.799e+01, 1.038e+01, 1.228e+02, ..., 2.654e-01, 4.601e-01,
        1.189e-01],
       [2.057e+01, 1.777e+01, 1.329e+02, ..., 1.860e-01, 2.750e-01,
        8.902e-02],
       [1.969e+01, 2.125e+01, 1.300e+02, ..., 2.430e-01, 3.613e-01,
        8.758e-02],
       ...,
       [1.660e+01, 2.808e+01, 1.083e+02, ..., 1.418e-01, 2.218e-01,
        7.820e-02],
       [2.060e+01, 2.933e+01, 1.401e+02, ..., 2.650e-01, 4.087e-01,
        1.240e-01],
       [7.760e+00, 2.454e+01, 4.792e+01, ..., 0.000e+00, 2.871e-01,
        7.039e-02]]), 'target': array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0,
       0, 0, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1, 1, 0, 1, 0, 0,
       1, 1, 1, 1, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 0,
       1, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 1, 1, 1, 1, 0, 1, 1, 0, 1,
	   .................................................................
       1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 1, 0, 1, 1,
       1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 0, 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 1]), 'frame': None, 'target_names': array(['malignant', 'benign'], ......................................,
       'mean smoothness', 'mean compactness', 'mean concavity',
       'mean concave points', 'mean symmetry', 'mean fractal dimension',
       'radius error', 'texture error', 'perimeter error', 'area error',
	   ................................................................
       'worst compactness', 'worst concavity', 'worst concave points',
       'worst symmetry', 'worst fractal dimension'], dtype='</u9'),>

In [3]:

breast_data.keys()

Out[3]:

dict_keys(['data', 'target', 'frame', 'target_names', 'DESCR', 'feature_names', 'filename', 'data_module'])

In [4]:

# data:문제데이터, 입력특성

# target: 정답데이터

# target_names: 정답데이터의 이름

In [5]:

# 정답데이터 확인

breast_data.target

Out[5]:

array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0,
       0, 0, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1, 1, 0, 1, 0, 0,
       1, 1, 1, 1, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 0,
       1, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 1, 1, 1, 1, 0, 1, 1, 0, 1,
       1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0, 1, 0,
       0, 1, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 0, 1, 1, 1, 1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 1, 1,
       1, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1, 1, 1, 0, 1, 1, 0, 0, 1, 0, 0,
       0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0,
       1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 1, 0, 0, 1, 0, 1, 1,
       1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 0, 1, 0, 0, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 0, 0,
       0, 1, 1, 1, 1, 0, 1, 0, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 0,
       0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0,
       1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 0, 0, 1, 1,
       1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 0,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0, 1, 0, 1, 1, 1, 1,
       1, 0, 1, 1, 0, 1, 0, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0,
       1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1,
       1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 1, 0, 1, 1,
       1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 0, 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 1])

In [6]:

# 정답데이터 확인

breast_data.target_names

# 0 == 'malignant'(악성), 1 == 'begin'(양성)

# 클래스의 개수: 2개 -> 이진분류

Out[6]:

array(['malignant', 'benign'], dtype='<U9')

In [7]:

breast_data.target_names

Out[7]:

array(['malignant', 'benign'], dtype='<U9')

In [8]:

# 문제, 정답 분리

# 문제 X, 정답 y

X = breast_data['data'] y = breast_data['target']

In [9]:

# train, test 분리

from sklearn.model_selection import train_test_split

# train_test_split(test_size = 0.2)

X_train, X_test, y_train, y_test = train_test_split(X,y, test_size = 0.2, random_state = 919)

In [10]:

print(X_train.shape, y_train.shape)

print(X_test.shape, y_test.shape)

(455, 30) (455,)
(114, 30) (114,)

 

모델링 과정

1. 신경망 구조 설계 (뼈대, 입력층, 중간층, 출력층)
2. 신경망 모델 학습/평가 방법 설정 (회귀, 분류)
3. 모델 학습
4. 모델 예측 및 평가

In [11]:

# 모델 생성을 위한 도구 불러오기

from tensorflow.keras.models import Sequential # 뼈대

from tensorflow.keras.layers import InputLayer, Dense

In [12]:

# 1. 모델 구조 설계

# 뼈대

model = Sequential()

 

# 입력층

model.add(InputLayer(input_shape = (30,))) # 입력층 (입력특성의 개수)

 

# 중간층(은닉층)

model.add(Dense(units = 16, activation = "sigmoid"))

model.add(Dense(units = 8, activation = "sigmoid"))

 

# 출력층 출력받고싶은 형태 지정( 이진분류, 0-1사이 확률값 하나)

model.add(Dense(units = 1, activation = "sigmoid"))

In [13]:

# 2. 모델 학습 및 평가 방법 설정

model.compile(loss = 'binary_crossentropy', # 오차 : 이진분류 -> binary_crossentropy

                        optimizer = 'SGD', # 최적화 알고리즘 (확률적 경사하강법)

                        metrics = ['accuracy']) # 평가방법(분류:정확도)

In [14]:

h1 = model.fit(X_train, y_train, validation_split = 0.2, epochs = 20)

Epoch 1/20
12/12 [==============================] - 1s 41ms/step - loss: 0.6615 - accuracy: 0.6456 - val_loss: 0.6753 - val_accuracy: 0.5934
Epoch 2/20
12/12 [==============================] - 0s 6ms/step - loss: 0.6585 - accuracy: 0.6456 - val_loss: 0.6716 - val_accuracy: 0.5934
Epoch 3/20
12/12 [==============================] - 0s 6ms/step - loss: 0.6538 - accuracy: 0.6456 - val_loss: 0.6701 - val_accuracy: 0.5934
Epoch 4/20
12/12 [==============================] - 0s 4ms/step - loss: 0.6515 - accuracy: 0.6456 - val_loss: 0.6695 - val_accuracy: 0.5934
Epoch 5/20
12/12 [==============================] - 0s 4ms/step - loss: 0.6500 - accuracy: 0.6456 - val_loss: 0.6687 - val_accuracy: 0.5934
Epoch 6/20
12/12 [==============================] - 0s 4ms/step - loss: 0.6479 - accuracy: 0.6456 - val_loss: 0.6680 - val_accuracy: 0.5934
Epoch 7/20
12/12 [==============================] - 0s 4ms/step - loss: 0.6462 - accuracy: 0.6456 - val_loss: 0.6678 - val_accuracy: 0.5934
Epoch 8/20
12/12 [==============================] - 0s 6ms/step - loss: 0.6448 - accuracy: 0.6456 - val_loss: 0.6662 - val_accuracy: 0.5934
Epoch 9/20
12/12 [==============================] - 0s 6ms/step - loss: 0.6434 - accuracy: 0.6456 - val_loss: 0.6658 - val_accuracy: 0.5934
Epoch 10/20
12/12 [==============================] - 0s 6ms/step - loss: 0.6422 - accuracy: 0.6456 - val_loss: 0.6664 - val_accuracy: 0.5934
Epoch 11/20
12/12 [==============================] - 0s 6ms/step - loss: 0.6413 - accuracy: 0.6456 - val_loss: 0.6648 - val_accuracy: 0.5934
Epoch 12/20
12/12 [==============================] - 0s 5ms/step - loss: 0.6402 - accuracy: 0.6456 - val_loss: 0.6644 - val_accuracy: 0.5934
Epoch 13/20
12/12 [==============================] - 0s 5ms/step - loss: 0.6396 - accuracy: 0.6456 - val_loss: 0.6645 - val_accuracy: 0.5934
Epoch 14/20
12/12 [==============================] - 0s 4ms/step - loss: 0.6387 - accuracy: 0.6456 - val_loss: 0.6634 - val_accuracy: 0.5934
Epoch 15/20
12/12 [==============================] - 0s 4ms/step - loss: 0.6382 - accuracy: 0.6456 - val_loss: 0.6637 - val_accuracy: 0.5934
Epoch 16/20
12/12 [==============================] - 0s 4ms/step - loss: 0.6372 - accuracy: 0.6456 - val_loss: 0.6621 - val_accuracy: 0.5934
Epoch 17/20
12/12 [==============================] - 0s 4ms/step - loss: 0.6365 - accuracy: 0.6456 - val_loss: 0.6625 - val_accuracy: 0.5934
Epoch 18/20
12/12 [==============================] - 0s 4ms/step - loss: 0.6358 - accuracy: 0.6456 - val_loss: 0.6617 - val_accuracy: 0.5934
Epoch 19/20
12/12 [==============================] - 0s 4ms/step - loss: 0.6352 - accuracy: 0.6456 - val_loss: 0.6615 - val_accuracy: 0.5934
Epoch 20/20
12/12 [==============================] - 0s 6ms/step - loss: 0.6343 - accuracy: 0.6456 - val_loss: 0.6605 - val_accuracy: 0.5934

 

In [15]:

# 결과 시각화 

plt.figure(figsize = (15, 5))

plt.plot(h1.history['loss'], label = 'train_loss')

plt.plot(h1.history['val_loss'], label = 'val_loss')

plt.legend() plt.show()

 

 

 

In [16]:

# 모델에 대한 전체적인 내부구조 확인 model.summary()

Model: "sequential"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 dense (Dense)               (None, 16)                496       
                                                                 
 dense_1 (Dense)             (None, 8)                 136       
                                                                 
 dense_2 (Dense)             (None, 1)                 9         
                                                                 
=================================================================
Total params: 641 (2.50 KB)
Trainable params: 641 (2.50 KB)
Non-trainable params: 0 (0.00 Byte)
_________________________________________________________________