[Deep Learning] Multi Variable Linear Regression

설명

• 퀴즈1, 퀴즈2, 중간고사 성적으로 기말고사 성적 예측하기

예제 코드

import tensorflow as tf

x1 = [73., 93., 89., 96., 73.]
x2 = [80., 88., 91., 98., 66.]
x3 = [75., 93., 90., 100., 70.]
Y = [152., 185., 180., 196., 142.]

w1 = tf.Variable(10.)
w2 = tf.Variable(10.)
w3 = tf.Variable(10.)
b = tf.Variable(10.)

learning_rate = 0.000001

for i in range(1000 + 1):
    with tf.GradientTape() as tape:
        hypothesis = w1 * x1 + w2 * x2 + w3 * x3 + b
        cost = tf.reduce_mean(tf.square(hypothesis - Y))

    w1_grad, w2_grad, w3_grad, b_grad = tape.gradient(cost, [w1, w2, w3, b])

    w1.assign_sub(learning_rate * w1_grad)
    w2.assign_sub(learning_rate * w2_grad)
    w3.assign_sub(learning_rate * w3_grad)
    b.assign_sub(learning_rate * b_grad)

    if i % 50 == 0:
        print("{:5} | {:12.4f}".format(i, cost.numpy()))

print("{:.4f} | {:.4f} | {:.4f} | {:.4f}".format(w1.numpy(), w2.numpy(), w3.numpy(), b.numpy()))

    0 | 5793889.5000
   50 |   64291.1484
  100 |     715.2902
  150 |       9.8462
  200 |       2.0152
  250 |       1.9252
  300 |       1.9210
  350 |       1.9177
  400 |       1.9145
  450 |       1.9114
  500 |       1.9081
  550 |       1.9050
  600 |       1.9018
  650 |       1.8986
  700 |       1.8955
  750 |       1.8923
  800 |       1.8892
  850 |       1.8861
  900 |       1.8829
  950 |       1.8798
 1000 |       1.8767
0.6694 | 0.6683 | 0.5595 | 9.8914

Matrix를 사용한 예제 코드

import numpy as np
import tensorflow as tf

data = np.array([
    [73., 80., 75., 152.],
    [93., 88., 93., 185.],
    [89., 91., 90., 180.],
    [96., 98., 100., 196.],
    [73., 66., 70., 142.],
], dtype=np.float32)

X = data[:, :-1]
y = data[:, [-1]]

W = tf.Variable(tf.compat.v1.random_normal([3, 1]))  # 입력 파라미터가 3개
b = tf.Variable(tf.compat.v1.random_normal([1]))  # 출력 결과가 1개

learning_rate = 0.000001

for i in range(2000 + 1):
    with tf.GradientTape() as tape:
        hypothesis = tf.matmul(X, W) + b
        cost = tf.reduce_mean(tf.square(hypothesis - y))

    W_grad, b_grad = tape.gradient(cost, [W, b])

    W.assign_sub(learning_rate * W_grad)
    b.assign_sub(learning_rate * b_grad)

    if i % 100 == 0:
        print("{:5} | {:10.4f}".format(i, cost.numpy()))

print(W.numpy())
print(b.numpy())

    0 |    37.5109
  100 |    19.2034
  200 |    19.1517
  300 |    19.1025
  400 |    19.0534
  500 |    19.0046
  600 |    18.9560
  700 |    18.9077
  800 |    18.8596
  900 |    18.8118
 1000 |    18.7643
 1100 |    18.7170
 1200 |    18.6700
 1300 |    18.6232
 1400 |    18.5767
 1500 |    18.5304
 1600 |    18.4843
 1700 |    18.4384
 1800 |    18.3926
 1900 |    18.3472
 2000 |    18.3019
[[ 0.02031493]
 [-0.79357684]
 [ 2.7809958 ]]
[-1.9658151]

참고

• https://youtu.be/tmfo04t664Q