1. Practice Project Overview
CSCE 4143: Date Mining
Yueyang Wang
Spring 2019

2. Data: Adult dataset

3. Description of dataset
Figure1: Boxplots of numeric attributes
Online Source:

4. Data Preprocessing: Remove records with unknown (
Data Preprocessing: Remove records with unknown (?) values from both train and test data sets

5. Data Preprocessing: Remove all continuous attributes

6. Q1.a Build a decision tree classifier (single tree) and report accuracy by class including (TP rate, FP rate, precision, recall, F1) on the test data.
Apply Weka

7. Q1.a Build a decision tree classifier (single tree) and report accuracy by class including (TP rate, FP rate, precision, recall, F1) on the test data.
Use Scikit-Learn

8. Q1.b  Build a naïve Bayesian classifier and report accuracy by class including (TP rate, FP rate, precision, recall, F1) on the test data.
Apply Weka

9. Q1.b  Build a naïve Bayesian classifier and report accuracy by class including (TP rate, FP rate, precision, recall, F1) on the test data.
Use ScikitLearn

10. Data Preprocessing: Use one-hot encoding to transform multi-domain categorical attribute
Apply Weka

11. Data Preprocessing: Use one-hot encoding to transform multi-domain categorical attribute

12. Data Preprocessing:  For each numerical attribute, use the mean value to transform into binary attribute
Use Python

13. Q2.a Build k-means clustering algorithm over train data with varied k values (3, 5, 10) based on your chosen distance function and report the centroids of the clusters
K=3
K=5
K=10

14. Q2.a Build k-means clustering algorithm over train data with varied k values (3, 5, 10) based on your chosen distance function and report the centroids of the clusters
Transform income (<=50K, >50K) to binary

15. Q2.b Use the last 10 records from test data and use kNN algorithm (with varied k values, 3, 5, 10)  to report the prediction accuracy.
K=3
K=5
K=10

16. Q2.b Use the last 10 records from test data and use kNN algorithm (with varied k values, 3, 5, 10)  to report the prediction accuracy.

17. Q3. Use the train datasets from step 2, build a SVM classifier and report the predicted accuracy of the test data. 
Apply Weka

18. Q3. Use the train datasets from step 2, build a SVM classifier and report the predicted accuracy of the test data. 

19. Q4. Use the train datasets from step 2, build a neural network classifier and report the predicted accuracy of the test data.
Apply Weka

20. Q4. Use the train datasets from step 2, build a neural network classifier and report the predicted accuracy of the test data.

21. Questions?