1 / 9

Machine Learning

Machine Learning. Lecture 11 Summary. Topics Covered. Topic 1: Introduction Learning and learning systems Design of learning systems - 5 steps approach: training sample collection/preparation, data representation, choose a learning model/paradigm, learning, testing

fitzgerald-oconnor
Télécharger la présentation

Machine Learning

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Machine Learning Lecture 11 Summary G53MLE | Machine Learning | Dr Guoping Qiu

  2. Topics Covered • Topic 1: Introduction • Learning and learning systems • Design of learning systems - 5 steps approach: training sample collection/preparation, data representation, choose a learning model/paradigm, learning, testing • Basic maths – vector, matrix, calculus G53MLE | Machine Learning | Dr Guoping Qiu

  3. Topics Covered • Topic 2: Neural networks • Perceptron – model operation, perceptron learning rule, decision boundary (surface), limitations, linearly separable classes • ADLINE – model operation, delta rule (gradient descent learning), batch mode, online mode • Multi-layer perceptron (MLP) - model operation, nonlinear unit (sigmoid function), principles of learning rule (back-propagation algorithm), model capability (solve linearly non-separable problems) • Generalization, over fitting, stopping criteria G53MLE | Machine Learning | Dr Guoping Qiu

  4. Topics Covered • Topic 3: Bayesian learning • Basic probability theory • Bayesian theorem • Estimation of probabilities • Maximum a posteriori (MAP) • Maximum Likelihood (ML) • Bayesian classifiers • Naïve Bayesian classifier G53MLE | Machine Learning | Dr Guoping Qiu

  5. Topics Covered • Topic 4: Instance based learning • K nearest neighbour classifier (K-NN) – feature space neighbourhood concept, classifier construction and operation, choose a suitable values of K G53MLE | Machine Learning | Dr Guoping Qiu

  6. Topics Covered • Topic 5: Clustering analysis • Basic concept of data clustering and why it is useful • How to do data clustering (K-means algorithm) – operation of the algorithm • Link between K-means algorithm and gradient descent (the concept of clustering cost function or objective function) • Limitations/weaknesses of the basic K-means algorithm – sensitive to initial cluster centres, local minima G53MLE | Machine Learning | Dr Guoping Qiu

  7. Topics Covered • Topic 6: Data processing and representation • Concepts of correlation and redundancy • Covariance matrix • Concepts of feature extraction/dimensionality reduction • Principle and application of Principal Component Analysis (PCA) G53MLE | Machine Learning | Dr Guoping Qiu

  8. Topics Covered • Topic 7: Support vector machines • Model operation (how it works) • Support vectors • Max margin classifier • Linear SVMs • Concept of Soft margin classification • Principle of Non-linear SVMs and the “Kernel Trick” G53MLE | Machine Learning | Dr Guoping Qiu

  9. Topics Covered • Topic 8: Decision tree learning • Information gain • Decision tree construction (design) – picking the root node, recursive branching G53MLE | Machine Learning | Dr Guoping Qiu

More Related