addinedu

blood fat

TensorFlow/Keras를 사용하여 몸무게와 나이 데이터로 혈중지방을 예측하는 선형 회귀 모델 구현. Sequential API를 활용한 간단한 신경망 모델 훈련 및 3D 시각화

Machine Learning
Python
Tensorflow

Keras 를 사용해서 몸무게, 나이 데이터를 가지고 혈중지방을 예측하는 모델을 만들어보기

In [1]:
# Blood Fat 데이터를 NumPy 배열로 변환
import numpy as np

# 원본 데이터 (Weight, Age, Blood fat content)
data = np.array([
    [84, 46, 354],
    [73, 20, 190],
    [65, 52, 405],
    [70, 30, 263],
    [76, 57, 451],
    [69, 25, 302],
    [63, 28, 288],
    [72, 36, 385],
    [79, 57, 402],
    [75, 44, 365],
    [27, 24, 209],
    [89, 31, 290],
    [65, 52, 346],
    [57, 23, 254],
    [59, 60, 395],
    [69, 48, 434],
    [60, 34, 220],
    [79, 51, 374],
    [75, 50, 308],
    [82, 34, 220],
    [59, 46, 311],
    [67, 23, 181],
    [85, 37, 274],
    [55, 40, 303],
    [63, 30, 244]
])

데이터 시각화

In [2]:
import matplotlib.pyplot as plt

xs = np.array(data[:,0], dtype=np.float32)
ys = np.array(data[:,1], dtype=np.float32)
zs = np.array(data[:,2], dtype=np.float32)

fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
ax.scatter(xs, ys, zs)
ax.set_xlabel('Weight')
ax.set_ylabel('Age')
ax.set_zlabel('Blood fat')
plt.show()
Notebook output
In [3]:
X = np.array(data[:, 0:2])
y = np.array(data[:, 2:])
y = y.reshape(-1, 1)

print(X.shape, y.shape)
(25, 2) (25, 1)
In [12]:
import tensorflow as tf

model = tf.keras.Sequential([
    tf.keras.layers.Input(shape=(2,)),
    tf.keras.layers.Dense(1)
])

model.compile(optimizer="rmsprop", loss="mse")
model.summary()
Model: "sequential_1"
Plain text view
Model: "sequential_1"

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Layer (type)                    ┃ Output Shape           ┃       Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ dense_1 (Dense)                 │ (None, 1)              │             3 │
└─────────────────────────────────┴────────────────────────┴───────────────┘
Plain text view
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Layer (type)                    ┃ Output Shape           ┃       Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ dense_1 (Dense)                 │ (None, 1)              │             3 │
└─────────────────────────────────┴────────────────────────┴───────────────┘

 Total params: 3 (12.00 B)
Plain text view
 Total params: 3 (12.00 B)

 Trainable params: 3 (12.00 B)
Plain text view
 Trainable params: 3 (12.00 B)

 Non-trainable params: 0 (0.00 B)
Plain text view
 Non-trainable params: 0 (0.00 B)

모델 학습

나이와 몸무게를 이용해 blood fat 을 추정하는 모델을 훈련

In [5]:
hist = model.fit(X, y, epochs=10000, verbose=0)
2026-01-03 20:14:54.124083: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:117] Plugin optimizer for device_type GPU is enabled.
In [6]:
import matplotlib.pyplot as plt

plt.plot(hist.history['loss'])
plt.title('Model loss')
plt.ylabel('Loss')
plt.xlabel('Epoch')
plt.show()
Notebook output

예측

몸무게와 나이를 주고 혈중지방을 예측해보기

In [7]:
model.predict(np.array([100, 44]).reshape(1, 2), verbose=0)
array([[373.1172]], dtype=float32)
In [8]:
model.predict(np.array([60, 25]).reshape(1, 2), verbose=0)
array([[220.09103]], dtype=float32)

훈련된 가중치와 bias 를 얻고 싶다면?

In [9]:
W_, b_ = model.get_weights()
print("Weight is", W_)
print("Bias is", b_)
Weight is [[1.1913167]
 [5.5459733]]
Bias is [9.962692]
In [10]:
x = np.linspace(20, 100, 50).reshape(50, 1)
y = np.linspace(10, 70, 50).reshape(50, 1)

X = np.concatenate((x, y), axis=1)
Z = np.matmul(X, W_) + b_
In [11]:
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
ax.scatter(xs, ys, zs)
ax.scatter(x, y, Z, color='r')
ax.set_xlabel('Weight')
ax.set_ylabel('Age')
ax.set_zlabel('Blood fat')
plt.show()
Notebook output
LinkedIn