I compared several income-classification methods in R as a learning project.
Result: The repository includes the original IncomeLevelPrediction.R script.
Tools
R
Logistic regression
Random forest
SVM
This is an exploratory learning project, not a validated accuracy benchmark or a deployed decision system.
Project gallery
Original historical presentation; exploratory results, not a validated benchmark. 26 images. Select an image to view it full size.
Original R study, slide 1: Title and authorOriginal R study, slide 2: ContentsOriginal R study, slide 3: Dataset and provenanceOriginal R study, slide 4: Data explorationOriginal R study, slide 5: Missing-value handlingOriginal R study, slide 6: Income and predictor explorationOriginal R study, slide 7: Occupation and marital-status explorationOriginal R study, slide 8: Exploratory plotsOriginal R study, slide 9: Race and gender explorationOriginal R study, slide 10: Train and test splitOriginal R study, slide 11: Logistic regression setupOriginal R study, slide 12: Updated logistic regressionOriginal R study, slide 13: Final logistic regression modelOriginal R study, slide 14: Variance inflation factorsOriginal R study, slide 15: Logistic regression evaluationOriginal R study, slide 16: Classification treeOriginal R study, slide 17: Tree pruningOriginal R study, slide 18: Tree evaluationOriginal R study, slide 19: Random forestOriginal R study, slide 20: Random-forest tuningOriginal R study, slide 21: Variable importanceOriginal R study, slide 22: Random-forest evaluationOriginal R study, slide 23: Linear SVMOriginal R study, slide 24: Linear SVM evaluationOriginal R study, slide 25: Model comparisonOriginal R study, slide 26: Discussion and conclusion
Technical details
Independent learning project · Historical R code
Different classification models can give different results on the same data.
Explore the input data and income categories.
Fit logistic regression, decision trees, random forests, and a linear SVM.
