Machine Learning I
6 Jan. 2013Tags: Hig
Machine Learning and Pattern Recognition I
Literature
- Required reading: Machine Learning and Data Mining: Introduction to Principles and Algorithms by Igor Kononenko and Matjaz Kukar (In library as "HIG, Pensum 006.3 Kon")
- Required reading: Selected papers
- Optional 1: R.O.Duda, P.E. Hart, and D.G. Storck: Pattern Classification. 2nd ed., Wiley, 2001.
- Optional 2: Sergios Theodoridis, Konstantinos Koutroumbas. “Pattern Recognition”, third edition. Academic Press.
- Coursera has at least two courses on machine learning.
Info
- Exam: 3 hours (70%)
- Graded homework: (3x 10%) (~15 hours each)
- Optional exercises: 7 (walk-through on Mondays) (~10 hours each)
- Lectures: 7 (Tuesdays)
- The goal of this course is to understanding methods, not knowing formulas. Know what methods are used for, where they do well and where they perform poorly. This subject will be salted with forensics principles.
Useful resources
- Andrew Ng on the future of robotics (neural networks)
- Videos: Probability (Khan Academy) (the basics)
- Video: Introduction to Machine Learning (good for starters)
- Video: Basics of probability and statistics (good for more advanced)
- Paper: Statistical Pattern Recognition: A Review
- Video: P vs NP problems
- Generative vs disciminative algorithms (Andrew Hg)
My lecture notes
- Lecture 1: Learning and intelligence and data analysis (chapter 1-2)
- Lecture 2: Performance of classifiers (chapter 3-4)]
- Lecture 3: Learning as search (chapter 5)
- Lecture 4: Feature selection and attribute quality measures (chapter 6-7)
- Lecture 5: Symbolic and statistical learning (chapter 9-10)
- Lecture 6: Cluster analysis (Chapter 10)
- Lecture 7: Artificial Neural Networks (Sparse learning) (Chapter 11)
My mind maps
Videos by mathematicalmonk
- Introduction
- Supervised learning
- Unsupervised learning
- Variations on supervised and unsupervised
- Generative vs discriminative
Videos by Andrew Ng (Stanford)
Inspiring videos
Snarveier
Flere fra skole