본문으로 바로가기

강의

R에서 tidymodels로 모델링하기

중급기술 수준

업데이트됨 2022. 7.

tidymodels로 머신러닝 워크플로를 효율적으로 구축하는 방법을 배우십시오.

무료로 강의 시작

RMachine Learning

4시간

16 동영상

59 연습 문제

4,950 XP

10,783

성취 증명서

수천 개 기업의 학습자들이 사랑하는

팀을 교육하시나요?

비즈니스용으로 체험해 보세요

강의 설명

Tidymodels는 Machine Learning 워크플로를 간소화하도록 설계된 강력한 R 패키지 모음입니다. 교차 검증을 위해 데이터셋을 분할하고, tidymodels의 recipes 패키지로 데이터를 전처리하며, Machine Learning 알고리즘을 미세 조정하는 방법을 배워 보세요. 모델 객체를 정의하고 모델링 워크플로를 만드는 핵심 개념도 익힙니다. 그런 다음, 주택 가격을 예측하고 직원의 이직 위험을 분류하는 실제 문제에 적용해 봅니다.

선수 조건

Modeling with Data in the Tidyverse

1

Machine Learning with tidymodels

In this chapter, you’ll explore the rich ecosystem of R packages that power tidymodels and learn how they can streamline your machine learning workflows. You’ll then put your tidymodels skills to the test by predicting house sale prices in Seattle, Washington.

The tidymodels ecosystem

Tidymodels packages

Creating training and test datasets

Distribution of outcome variable values

Linear regression with tidymodels

Fitting a linear regression model

Exploring estimated model parameters

Predicting home selling prices

Evaluating model performance

Model performance metrics

R squared plot

Complete model fitting process with last_fit()

2

Classification Models

Learn how to predict categorical outcomes by training classification models. Using the skills you’ve gained so far, you’ll predict the likelihood of customers canceling their service with a telecommunications company.

Classification models

Data resampling

Fitting a logistic regression model

Combining test dataset results

Assessing model fit

Calculating metrics from the confusion matrix

Evaluating performance with yardstick

Creating custom metric sets

Visualizing model performance

Plotting the confusion matrix

ROC curves and area under the ROC curve

Automating the modeling workflow

Streamlining the modeling process

Collecting predictions and creating custom metrics

Complete modeling workflow

3

Feature Engineering

Find out how to bake feature engineering pipelines with the recipes package. You’ll prepare numeric and categorical data to help machine learning algorithms optimize your predictions.

Feature engineering

Exploring recipe objects

Creating recipe objects

Training a recipe object

Numeric predictors

Discovering correlated predictors

Removing correlated predictors with recipes

Multiple feature engineering steps

Nominal predictors

Applying step_dummy() to predictors

Ordering of step_*() functions

Complete feature engineering pipeline

Complete modeling workflow

Feature engineering process

Model training and prediction

Model performance metrics

4

Workflows and Hyperparameter Tuning

Now it’s time to streamline the modeling process using workflows and fine-tune models with cross-validation and hyperparameter tuning. You’ll learn how to tune a decision tree classification model to predict whether a bank's customers are likely to default on their loan.

Machine learning workflows

Exploring the loans dataset

Specifying a model and recipe

Creating workflows

Estimating performance with cross validation

Measuring performance with cross validation

Cross validation with logistic regression

Comparing model performance profiles

Hyperparameter tuning

Setting model hyperparameters

Random grid search

Exploring tuning results

Selecting the best model

Finalizing a workflow

Training a finalized workflow

Congratulations!

R에서 tidymodels로 모델링하기

강의
완료

수료증 획득

LinkedIn 프로필, 이력서 또는 CV에 이 인증서를 추가하세요
소셜 미디어와 성과 평가에서 공유하세요지금 등록

19백만 명 이상의 학습자와 함께 R에서 tidymodels로 모델링하기을(를) 시작하세요!

DataCamp for Mobile을 통해 데이터 분석 능력을 향상시키세요.

모바일 강좌와 매일 5분 코딩 챌린지를 통해 이동 중에도 학습 효과를 높이세요.