NOTE: ID/PWD to access course materials will be distributed in the first class
| Week |
Date
|
Contents |
Reading List |
Download | Assigned |
DUE |
| 1 | 8/26
|
Logistics
|
|
logistics | hw0 |
|
| | 8/28 | Introduction to ML | DM4 Ch 1, CIML Ch 1.2, MMDS Ch 1 | intro, lab-personality | | hw0 | | 2 | 9/2
| NO CLASS (Labor Day) |
|
|
|
| | | 9/4 | Inputs and Outputs: nominal, ordinal, interval, ratio | DM4 Ch 2-3, IDM2 Ch 2.1
| in-out, lab-iris |
|
| | 3 | 9/9 | Jupyter Notebook Lab | Weka, Jupyter | lab-jupyter, lab-heart-disease (azure) |
|
| | | 9/11 | Distance: euclidean, minkowski, mahalanobis, mutual information | IDM2 Ch 2.4 | similarity | hw1 |
| | 4 | 9/16 | Decision tree: hunt's algorithm, gini index, entropy, information gain
| CIML Ch 1, IDM2 Ch 3.3, DM4 Ch 4.3
| decision-tree |
|
| | | 9/18 | 1R and Naive Bayes Classification | CIML Ch 9.3, IDM2 Ch 4.4, DM4 Ch 4.2 |
| proj1
| hw1 (9/22) | | 5 | 9/23 | Evaluation: confidence interval, cross-validation, bootstraping, hyperparameter, model comparison | DM4 Ch 5, CIML Ch 2.5-2.7, Ch 5.5-5.7 |
|
|
| | | 9/25 | Evaluation: loss function, cost-sensitive learning, confusion matrix, various error measures | DM4 Ch 5, CIML Ch 1.4 |
| hw2 |
| | 6 | 9/30 | Limits of Learning: underfitting, overfitting | CIML Ch 2.4, IDM2 Ch 3.4 |
|
|
| | | 10/2 | Association rules: apriori, fp-growth | DM4 Ch 4, IDM2 Ch 5
|
|
| hw2 (10/6) | | 7 | 10/7 | Weka & Jupyter Notebook Lab |
|
|
|
| | | 10/9 | Linear Models |
|
|
| proj1 (10/13) | | 8 | 10/14 | Linear Models: regression, perceptron, winnow | DM4 Ch 4, CIML Ch 4, MMDS Ch 12.2
|
|
|
| | | 10/16 | MIDTERM |
|
|
|
| | 9 | 10/21 | Clustering: partitional vs. hierarchical, k-means vs. agglomerative, inter-cluster distances, incremental | DM4 Ch 7, CIML Ch 3, MMDS Ch 7 |
| proj2 |
| | | 10/23 | Project #2 Discussion |
|
|
|
| | 10 | 10/28 | Support Vector Machine (SVM): support vectors, hyperplane, margin, slack penalty, hinge loss
| DM4 Ch 7, CIML Ch 7.7, MMDS Ch 12.3
|
|
|
| | | 10/30 | Support Vector Machine (SVM): gradient decent, SGD, kernels, SMO (weka)
| DM4 Ch 7, CIML Ch 11.5-11.6, MMDS Ch 12.3
|
|
|
| | 11 | 11/4 | Project #2 Discussion |
|
|
|
| | | 11/6 | Feature Selection: filter, wrapper, embedded, forward stepwise, backward elimination, hill-climbing
| DM4 Ch 8, CIML Ch 5.1-5.4
|
|
|
| | 12 | 11/11 | Dimensionality Reduction: PCA, SVD | DM4 Ch 8, CIML Ch 3.5, Ch MMDS 11 |
|
|
| | | 11/13 | Ensemble Learning: bagging, randomization, boosting | DM4 Ch 12, CIML Ch 13, IDM2 Ch 4.10
|
|
|
| | 13 | 11/18 | Deep Learning | IDM2 Ch 4.7-4.8, DM4 Ch 10 |
|
|
| | | 11/20 | Deep Learning
| IDM2 Ch 4.7-4.8, DM4 Ch 10 |
|
|
| | 14 | 11/25
| NO CLASS (Thanksgiving Holiday) |
|
|
|
| | | 11/27 | NO CLASS (Thanksgiving Holiday) |
|
|
|
| | 15 | 12/2 | Project #2 Discussion |
|
|
|
| | | 12/4 | ML applications
| DM4 Ch 12, CIML Ch 2.7
|
|
| proj2
| | 16 | 12/9 | Presentation
|
|
|
|
| | | 12/11 | Presentation
|
|
|
|
| | 17 |
| FINAL EXAM
| |
| | |
|
|