The course is for the student to obtain an introductory understanding of algorithms and applications of machine learning. Topics include:
- Decision theory;
- Parametric models;
- Supervised learning(classification and regression);
- Unsupervised learning(clustering, mixture models, principal component analysis);
- Bayesian methods.
The objective of this course are to let student to:-
- understantd the motivations and principles for building adaptive systems based on empirical data, and how machine learning relates to the broader field of artificial intelligence; and
- formulate problems associated with domains specific data(e.g.: image classification, document clustering) in terms of abstract models of machine learning; and
- implement solutions to machine learning problems using tools such as Matlab or Octave, apply numerical optimization algorithms.
Through this course, I practically obtained the following techniques and principles regarding machine learning:
1. Data Pre-processing
1.1 Normalization
1.2 Standardization
1.3 Data Clearning
1.4 Data Selection(SLR w/ L1-regularization)
2. Classification Model
2.1 Perceptron, Adaline
2.2 Logistics Regression, Softmax Regression
2.3 Linear SVM, kernel SVM
2.4 KNN
2.5 Decision Tree, Random Forest
2.6 Ensemble Learning (Voting by Majority, Bagging, Adaboost, etc)
2.7 Sentiment Analysis
3. Regression Model
3.1 Simple Linear Regression
4. Dimensionality Reduction
4.1 Principal Component Analysis, PCA
4.2 Local Linear Embedding, LLE
5. Evaluation and Tuning
5.1 K-Fold Cross-Validation
5.2 Hyperparameter Grid Search
5.3 Confusion Matrix
5.4 Receiver Operating Characteristics(ROC)
5.5 Area under curve(AUC)
6. Convolutional Neural Network(CNN)
6.1 Convolution layer