Machine Learning Course

Rui Xia
School of Computer Science & Engineering
Nanjing University of Science & Technology

Course Information (including contacts, syllabus, assessment, references, etc.)

Contents

• An Introduction to Machine Learning [slides]
• Linear Regression [slides]
• Logistic Regression & Softmax Regression [slides]
• Perceptron Algorithm [slides]
• Review of Linear Models [slides]
• Multi-layer Forward Neutral Network & Back Propagation [slides]
• Concept Revisit (Generative vs. Discriminative, Overfitting and Regularization) [slides]
• Naive Bayes Model [slides]
• Support Vector Machines [slides]

Projects

1. Implement linear regression (Analytic, GD) for Nanjing housing price prediction [page 12 of slides]
2. Implement logistic regression (GD, SGD, Newton) and softmax regression for admit/non-admit binary classification [page 15 of slides]
3. Implement softmax regression (GD, SGD) for the same problem in Practice 2 and compare it with logistic regression [page 23 of slides]
4. Implement Perceptron and multi-class Perceptron for the same problem in Practice 2, and compare them with logistic regression and softmax regression (SGD) respectively [page 10 of slides]
5. Implement 3-layer Forward Neutral Network for the same problem in Practice 2 with 5-fold cross validation, based on 1) self-coding, and 2) Tensorflow resepcitively, and compare them [page 15 of slides]
6. Implement naïve Bayes algorithm with Multinomial event model and Multi-variate Bernoulli model, and run the algorithm based on the training & testing data given in [page 18 of slides]

Last updated by Rui Xia on 2019-1-9