Homework _10 Spring 11

Homework _10 Spring 11 - 1 BME 6360 Homework #10 Spring...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 1 BME 6360 Homework #10 Spring 2011 Neural Engineering Prof. Wheeler The goal of this exercise is to practice basic pattern recognition methodology with biological signals. The report should consist of: (a) The output of the classplt function operating on the spikes separated by peak amplitude, annotated so as to indicate which class is which. (b) The same output for the principal component technique (c) The cluster plot for the principal component technique annotated so as to indicate which class is which (d) The same cluster plot as in (c) but with the confidence contours for each class. You will need the following files: spikes.dat, ident.dat, separate.m, classplt.m, classify.m, confused.m, cluster.m The data file hw10soln.mat has been included to help you if you get stuck. Try not to use it unless needed, but it is there to help you as the manipulations here are complex. It includes sp1, sp2, sp3, sp4 %separated spikes from part 1 peak, valley, pkpk % peak and valley values from part 2 pk1m, pk2m, pk3m, pk4m, pkm, % mean peak values from part 2 val1m, val2m, val3m, val4m, valm % mean valley values from part 2 pkpk1m, pkpk2m, pkpk3m, pkpk4m, pkpkm % mean peak-to-peak values from part 2 wave1m, wave2m, wave3m, wave4m, wavem % mean waveforms from parf 3 grandmean, gm, vary, covmat % covariance calculations from part 4 pc1m, pc2m, pc3m, pc4m, pcm % mean principal component values from part 4 Note: the most common data manipulation error you will make in this assignment is creating row vectors when you need column vectors, and vice versa. Pay close attention to the use of the transpose operator. If you have errors, the first thing to try is to re-execute the function with the transpose of the variable. 1. Preliminary Examination of the Data You have been given a file of action potential data ( spikes.dat ), as well as a file of class identifications ( ident.dat ) which were created by a human observer using a computer display. After you load these files, the matrix spikes will have 82 rows, one for each spike, and 35 columns, which index time. The column vector ident consists of integers equal to 1, 2, 3 or 4, depending on whether the corresponding row of spikes is an action potential belonging to the 1st, 2nd, 3rd, or 4th class. Plot the spikes to see the raw data. A program has been written which will separate the spike matrix into individual matrices according to the class identifier. Separate the spikes and plot the individual classes to investigate how well the human classifier did. There are some obvious errors, which is typical of humanly classified data in many fields: it is very difficult to get accurate training data. load spikes.dat % the resulting data matrix is spikes load ident.dat % the resulting class identifier vector is ident plot(spikes'); % you'll need to plot the transpose 2 As you plot each class of spikes, write down the approximate peak and valley amplitudes....
View Full Document

This note was uploaded on 01/23/2012 for the course EEL 6502 taught by Professor Principe during the Spring '08 term at University of Florida.

Page1 / 5

Homework _10 Spring 11 - 1 BME 6360 Homework #10 Spring...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online