Skip to Content
AboutMachineLearning

MachineLearning

 
pagePassword: True

준비중..

머신러닝은 인공지능의 한 분야로서 여전히 유효하게 사용 및 발전되고있습니다.

머신러닝 기본

  • 머신러닝 개요: 머신러닝의 정의, 종류, 응용 분야
  • 지도학습: 선형 회귀, 로지스틱 회귀, 의사결정나무, 랜덤 포레스트, SVM, k-NN
  • 비지도학습: 클러스터링, 차원 축소, PCA
  • 평가 지표: 정확도, 정밀도, 재현율, F1 점수, ROC-AUC 심화 학습
  • 앙상블 학습: 배깅, 부스팅, 스태킹
  • 특징 선택 및 엔지니어링: 데이터 전처리, 특징 선택 기법
  • 하이퍼파라미터 튜닝: 그리드 서치, 랜덤 서치, 베이지안 최적화
  • 모델 해석 및 평가: 모델 해석 기법, 오버피팅과 언더피팅 대응 방법 실습 프로젝트
  • 실제 데이터셋을 이용한 머신러닝 모델 구축 및 평가
  • 다양한 분야의 데이터셋을 활용한 실습 (예: 금융, 의료, 소셜 미디어 분석) PCA

Yellowbrick: Machine Learning Visualization https://www.scikit-yb.org/en/latest/ 

https://pycaret.org/ 

→ Automate ML workflows with this low-code library.

https://imbalanced-learn.org/stable/ 

https://modin.readthedocs.io/en/stable/ 

→ Boost Pandas’ performance up to 70x by modifying the import.

https://shap.readthedocs.io/en/latest/ 

https://facebook.github.io/prophet/ 

https://pypi.org/project/parallel-pandas/ 

https://www.featuretools.com/ 

→ Automated feature engineering for ML models.

Lazy Predict https://lazypredict.readthedocs.io/en/latest/# 

https://rasbt.github.io/mlxtend/ 

https://vaex.io/ 

High performance package for lazy Out-of-Core DataFrames

https://pypi.org/project/sweetviz/ 

In-depth EDA report in two lines of code.

→ Leverage the power of PyTorch with the elegance of sklearn.

→ Efficient algorithms for similarity search and clustering dense vectors.

https://www.statsmodels.org/stable/index.html 

https://pypi.org/project/pandas-profiling/ 

→ Generate a high-level EDA report of your data in no time.

Streamlit

Category-encoders

DuckDB

PandasML

Pytest → An elegant testing framework to test your code.

Numexpr → Parallelize NumPy to all CPU cores for 20x speedup

CSV-Kit → Explore, query and describe CSV files from terminal.

PivotTableJS → Drap-n-drop tools to group, pivot, plot dataframe.

Faker → Generate fake yet meaningful data in seconds.

Icecream → Don’t debug with print(). Use icecream.

Pyforest → No need to write imports. Automatic package import

PySnooper → Profile your code. Track new variables, and their updates.

Sidetable Supercharge Pandas’ value_counts() method.

https://www.blog.dailydoseofds.com/p/free-daily-dose-of-data-science-pdf 

ExcelOrange