Basic Concept of Machine Learning

• What is the data science?
• Why use machine learning?
• Traditional learning vs Mahcine learning
• Types of machine learning
• Conditions for the better machine learning system
• Testing and evaluating the machine learning system

Introduction

• ML WorkFlow Introduction
• Understanding Big Picture of Machine Learning

Data Collection & Preparation

• Structured Data vs Unstructured Data
• Sample, Feature, Target
• Supervised Learning vs Unsupervised Learning
• Regression vs Classification
• Data Cleansing

Preprocessing

• 데이터 유효성 탐지
• 데이터 통계 요약 분석
• EDA ( Binning / Smoothing )
• 데이터 전처리 방법

Modeling & Evaluation

• Modeling
• Cross Validation
• Overfitting
• 과적합이란?
• Parameter Tunning

What is the Algorithm? 

• 알고리즘이란?
• 예측함수 알고리즘의 산출물

Algorithms in DAVinCI LABS

• Ridge Regression
• Decision Tree ( 의사결정나무 )
• Random Forest
• GBM ( Gradient Boosting Machine )
• Neural Network ( 인공신경망 )
• Stabilized Deep Neural Network
• Support Vector Machine
• Ensembles ( 앙상블기법 )

Data Split

• Training set, Validation set, Test set
• 불균형 데이터 (Over/Under Sampling)
• Cross Validation (교차검증)

모델 성능 평가지표

• 숫자형 데이터 (MSE / MAE / GINI / KS / PSI / AUPRC / AIC)
• 범주형 데이터 (Accuracy / Precision / Recall / FI Score)

모델 성능 등급차트

• 숫자형 데이터
• 등급차트의 활용 예

Small Dataset Problem

Imbalanced Dataset Problem

Topic 1: Predict house price

Topic 2: Predict survival on the Titanic