๋ณธ๋ฌธ ๋ฐ”๋กœ๊ฐ€๊ธฐ

Study_note(zb_data)/Machine Learning

์Šคํ„ฐ๋””๋…ธํŠธ (Cost Function)

๐Ÿ“Œ Cost Function์€ ์ตœ์†Œํ™” ํ•  ์ˆ˜ ์žˆ๋Š” ๊ฒƒ์ด ์ข‹๋‹ค.

์ถœ์ฒ˜ : ์ œ๋กœ๋ฒ ์ด์Šค ๋ฐ์ดํ„ฐ ์Šค์ฟจ
์ถœ์ฒ˜ : ์ œ๋กœ๋ฒ ์ด์Šค ๋ฐ์ดํ„ฐ ์Šค์ฟจ
์ถœ์ฒ˜ : ์ œ๋กœ๋ฒ ์ด์Šค ๋ฐ์ดํ„ฐ ์Šค์ฟจ

๐Ÿ“Œ ์ฝ”๋“œ๋กœ ํ•œ ๋ฒˆ ๊ตฌํ•ด๋ณด์ž!

import numpy as np
a = np.poly1d([1, 1])
b = np.poly1d([1, -1])
a,b 
>>>>
(poly1d([1, 1]), poly1d([ 1, -1]))
# (x+1) * (x-1) = x^2 -1
a*b
>>>>
poly1d([ 1,  0, -1])
# x^2์˜ ๊ณ„์ˆ˜ -> 1 / x์˜ ๊ณ„์ˆ˜ -> 0 / ์ƒ์ˆ˜ํ•ญ -> -1
np.poly1d([2, -1])**2 + np.poly1d([3, -5])**2 + np.poly1d([5, -6])**2
>>>>
poly1d([ 38, -94,  62])
import sympy as sym
theta = sym.Symbol('th')
diff_th = sym.diff(38*theta**2-94*theta+62 , theta) # ์•ž์˜ ์‹์„ theta๋กœ diff(๋ฏธ๋ถ„) ํ•œ๋‹ค
diff_th
>>>>
76th−94
# 76th-94 = 0 ์œผ๋กœ ๊ทน์†Ÿ๊ฐ’์„ ๊ตฌํ•  ์ˆ˜ ์žˆ๋‹ค.

๐Ÿ“Œ ์‹ค์ „ ๋ฐ์ดํ„ฐ๋ฅผ ๋‹ค๋ฃฐ ๋•Œ๋Š” ๋‹ค๋ฅด๋‹ค..

- ์‹ค์ œ ๋ฐ์ดํ„ฐ๋Š” ๋„ˆ๋ฌด ๋ณต์žกํ•˜์—ฌ ์œ„ ์ฒ˜๋Ÿผ ์†์œผ๋กœ ํ’€๊ธฐ๊ฐ€ ์–ด๋ ต๋‹ค

์ถœ์ฒ˜ : ์ œ๋กœ๋ฒ ์ด์Šค ๋ฐ์ดํ„ฐ ์Šค์ฟจ
์ถœ์ฒ˜ : ์ œ๋กœ๋ฒ ์ด์Šค ๋ฐ์ดํ„ฐ ์Šค์ฟจ