일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 5 | ||
6 | 7 | 8 | 9 | 10 | 11 | 12 |
13 | 14 | 15 | 16 | 17 | 18 | 19 |
20 | 21 | 22 | 23 | 24 | 25 | 26 |
27 | 28 | 29 | 30 |
- local prediction
- os
- Aegis
- 언리얼 엔진
- Unreal Engine
- ability task
- UI
- 게임 개발
- 유니티
- attribute
- gameplay effect
- Replication
- C++
- widget
- CTF
- nanite
- stride
- gas
- Multiplay
- listen server
- photon fusion2
- MAC
- gameplay ability system
- animation
- 언리얼엔진
- gameplay tag
- unity
- 게임개발
- 보안
- rpc
- Today
- Total
목록2025/04/12 (10)
Replicated

def sigmoid(z): return 1/( 1 + np.exp(-z) )import matplotlib.pyplot as pltx = np.linspace(-10, 10, 400)y = sigmoid(x)plt.plot(x, y, label="Sigmoid Function")이렇게 그려짐 가설함수- 시그모이드 함수의 z -> 가중치와 피쳐의 선형 결합임- 피쳐 값들을 x 벡터, 가중치 값들은 세타로 입력def hypothesis_function(x, theta): z = (np.dot(x,theta)) return sigmoid(z)벡터 dot 시 알아서 내적임 비용함수def compute_cost(x, y, theta): m = y.shape[0] J = (-1.0 / m) * (y.T...

모델을 평가하는 성능 지표들- 회귀(regression): MAE, MSE, RMSE, SSE- 분류(classification): 정확도, 정밀도, 민감도, F1 스코어, ROC 커브, 리프트 차트- 클러스터링(clsutering): DBI, 엘보우 메서드, 실루엣 계수 추가적인 성능 지표- 모델의 경제성- 모델이 쓰는 데이터 양- 모델이 용량이 작은 컴퓨터에서도 돌아가나? 분류 문제의 성능지표혼동행렬(confusion matrix)- 예측값이 실제값 대비 얼마나 잘 맞는지 2x2 행렬로 표햔 예측값(prediction)실제값(actual class) 101True PositiveFalse Negative0False PositiveTrue NegativeTrue Positive(TP) : 예측값, 실..