★ Complete H1 Summary Notes ★ COMP30027: Machine Learning

Description

This note set covers the full subject of COMP30027 – Machine Learning for the 2022 Semester 1 Curriculum, including all covered concepts. The content of these notes were derived from the textbook, lecture notes, lecture recordings, workshop/tutorial content and examples. These notes include: - Multiple diagrams, tables and highlighted equations that are all clearly set out - Comprehensible diagrams and examples for difficult concepts - Summary tables that compare concepts taught, including advantages and disadvantages - Clear separation and format of concepts into topics and sections - Hyperlinked bookmarks of sections that can be easily navigated if opened with a PDF reader software The main topics covered include (but are not limited to): - Machine Learning Introduction (terminology, variables etc.) - Probability & Naïve Bayes (incl. probabilistic theorems, distributions, Naïve Bayes, data type conversion, KDE etc.) - Classifier Performance (incl. performance, test data selection, evaluation metrics, multiclass evaluation, generalisation, bias & variance etc.) - Decision Trees, K-NN, SVM (incl. one-R, decision trees, attribute selection, ID3 algorithm, instance-based learning, K-NN, SVM etc.) - Interpretation & Visualisation (incl. interpreting models, parameters, visualisation etc.) - Regression (incl. linear regression, optimization, evaluation, logistic regression etc.) - Further Classification (incl. classifier combination, feature selection & wrappers, filtering, HMM etc.) - Neural Networks (incl. neural networks, perceptron, deep learning & CNN etc.) - Unsupervised Learning (incl. clustering, GMM & EM etc.) - Semi-supervised Learning (incl. self-training, active learning, data augmentation) These notes helped me obtain an H1 in COMP30027: Machine Learning. They are perfect for both cramming and continuous study throughout the subject. The topics are covered in mostly the same order as they are covered in lectures, and is succinct and compact. All information included is relevant and important for the exam.