Study Guide 2015-2016

SGN-41006 Signal Interpretation Methods, 4 cr

Person responsible

Heikki Huttunen

Lessons

Implementation 1: SGN-41006 2015-01

Study type P1 P2 P3 P4 Summer
Lectures
Excercises

 

  		 4 h/week
 2 h/week

 

 

Lecture times and places: Tuesday 10 - 12 TB111 , Friday 10 - 12 TB111 , Tuesday 10 - 12 TB111

Requirements

Exam, homeworks and exercises.
Completion parts must belong to the same implementation

Learning Outcomes

Students understand principles of selected statistical, pattern recognition and machine learning approaches for interpreting signals. Student can apply the methods to real problems.

Content

Content Core content Complementary knowledge Specialist knowledge
1. Fundamentals: Probabilistic formulation of decision theory; The Bayes classifier; Naive Bayes Classifier  Detection theory; Receiver Operating Characteristics; Neyman-Pearson decision rule   
2. Statistical Signal Processing: Estimation theory; Maximum likelihood; Estimation of signal parameters (e.g., phase, amplitude and frequency).     
3. Linear models: regression and classification, support vector machines, logistic regression, regularisation.  Robust estimators: RANSAC, M-estimator.   
4. Performance evaluation, cross-validation, bootstrapping     
5. Modern tools (Random forests, Bagging, Boosting, Stacking, Deep Learning)     
6. Implementations: 1) scikit-learn, 2) pylearn2, 3) caffe, 4) Matlab Statistics toolbox     

Instructions for students on how to achieve the learning outcomes

Accepted exercises and homeworks. Final exam.

Assessment scale:

Numerical evaluation scale (1-5) will be used on the course

Partial passing:

Completion parts must belong to the same implementation

Study material

Type Name Author ISBN URL Additional information Examination material
Book   Elements of Statistical Learning: Data Mining, Inference, and Prediction   T. Hastie, R. Tibshirani and J. Friedman       Main course textbook. Freely available at http://www-stat.stanford.edu/~tibs/ElemStatLearn/   Yes   
Book   Pattern Recognition and Machine Learning   Christopher M. Bishop   0-387-31073-8     Additional reading   No   
Book   Statistical Pattern Recognition   Andrew R. Webb and Keith D. Copsey       Additional reading   No   

Prerequisites

Course Mandatory/Advisable Description
SGN-13000 Johdatus hahmontunnistukseen ja koneoppimiseen Mandatory   1
SGN-13006 Introduction to Pattern Recognition and Machine Learning Mandatory   1

1 . Either course SGN-13000 or SGN-13006 or equivalent knowledge.



Correspondence of content

Course Corresponds course  Description 
SGN-41006 Signal Interpretation Methods, 4 cr SGN-2556 Pattern Recognition, 5 cr  

Last modified 27.01.2015