[ML] 목차

목차

 

• 00. Overview
• 01. Naive Bayes
• 02. Text Data
• 03. Evaluation
• 04. Decision Tree
• 05. Support Vector Machine (SVM)
• 06. Logistic Regression
• 07. Neural Network
• 08. Deep Neural Network
• 09. Feature Engineering
• 10. Feature Selection
• 11. Feature Extraction

00. Overview

• AI, ML, DL
• 1) 특성 2) 타겟값 3) 알고리즘 4) 성능측정방법

01. Naive Bayes

• 조건부확률과 베이즈정리
• Naive Bayes 동작원리

02. Text Data

• 자연어처리 NLP
• 텍스트전처리
• 텍스트 특성

03. Evaluation

• 성능평가척도
• Confusion Matrix
• Accuracy, Precision, Recall, Specificity
• 성능척도간의 Trade-off
• 종합성능척도 → F1 Score, AUC(ROC Curve)

04. Decision Tree

• 트리분할측정척도 →지니불순도, IG(정보이득)
• 트리분할알고리즘-CART

05. Support Vector Machine (SVM)

• SVM
• Soft SVM
• 비선형문제 → 특성공간변환

06. Logistic Regression

• Logistic Regression
• Cost Function
• Gradient Descent
• Stochastic Gradient Descent (SGD)
• Mini-Batch Gradient Descent
• Regularization
• Softmax Fuction

07. Neural Network

• 퍼셉트론
• 역전파

08. Deep Neural Network

• 과적합 대응 →Drop out, Regularization, Early Stopping
• Vanishing/Exploding Gradient 대응 기법
  → Activation Func, Weight Initialization,Batch/Layer Normalization, Gradient Clipping, 경사하강법 개선기법

09. Feature Engineering

• Example

10. Feature Selection

• Curse of Dimensionality 차원의 저주
• Dimensionality Reduction 차원감소
• Correlation 상관관계
• Pearson Correlation
• Mutual Information 상호정보
• RFE (Recursive Feature Elimination)
• L1 정규화

11. Feature Extraction

• 특성추출
• 주성분분석 PCA
• LDA