Workflow
- Question or problem defintion (資料前處裡)
- Acquire training and testing data (資料前處裡)
- Wrangle, prepare, cleanse the data (資料前處裡)
- Analyze, identify patterns, and explore the data (資料前處裡)
- Model, predict and solve problem (Machine Learning)
- Visualize, report, and present the problem solving steps and final solution (Machine Learning)
- Supply or submit the results (Machine Learning)
7 Cs
- Classifying → Classify or categorize samples
- Correlating → Find features contribute significantly
- Converting → Depends on the problems
- Completing → Data preparation
- Correcting → Discard features significantly skew the result
- Creating → Create new features
- Charting → Visualization right plots
Kaggle 教學