Learning the Python ecosystem for machine learning

Summary

• Published 2025-01-30 · slug machine-learning-first-steps · category "AI and machine learning studies / ML foundations and practice" · source article recorded in Korean and English.
• A learning record of the roles of NumPy, Pandas, and Scikit-learn, and their relationships, written for future reference while starting machine learning.
• Code examples cover NumPy ndarrays and broadcasting, including the difference between shapes (100,) and (100, 1); Pandas DataFrames and groupby aggregation; and Scikit-learn's consistent .fit() and .predict() interfaces.
• Reflection: NumPy's concepts were the hardest of the three to grasp. Scikit-learn was quicker to learn because its interfaces are consistent.
• Next goals: study ensembles such as Random Forest and XGBoost, and enter a Kaggle competition.

Mentioned

• Kaggle: the competition platform mentioned as a next goal.

Connections

• Writing
• Concepts: XGBoost · i18n and multilingual support · Running a technical blog · Ensembles