we will consider use of PCA for simple dimensionality reduction,
i.e., determining the signal subspace when there are more observations than the underlying latent
variables—signals. The main assumption here is that both the noise and signals are independent and
identically distributed Gaussians, however the the signals are correlated among themselves while the
noise components are not, they are all uncorrelated—and hence independent as well. Assume there
are five signal components and five noise components, and generate 1000 samples of each. You can
use knee point detection function available in Matlab [login to view URL]
fileexchange/35094-knee-point to determine the order of signal subspace, i.e., number of signal
components. Generate two different correlation profiles for signals such that for one of those knee point
detection will work well but will not be able to determine the true order of five for the second case.
Make sure that all signal and noise components have the same variance, which is unit variance.
I have worked a lot with PCA and similar algorithms. Please see my reviews for more information.
Relevant Skills and Experience
MATLAB, Signal processing, C, C++, Verilog, VHDL, Python and so on.
Proposed Milestones
$399 USD - milestone
Hello! I have a wide experience with signal processing and transformation application and a good math foundation behind it, so I feel I can help you with this.
Relevant Skills and Experience
Signal processing, PCA, MATLAB
Proposed Milestones
$315 USD - All
I can do this project using Matlab.
Relevant Skills and Experience
I am expert in Signal/image processing, machine learning and pattern recognition.
PhD University of Antwerpen, Belgium
MS (EE) University of Southern California, Los Angeles,USA
Proposed Milestones
$250 USD - Final Milestone after commpletion
Send me the description pdf/txt file of your homework.
Hello, how are you? I have read the details provided, but please contact me so that we can discuss more on the project. I don't outsource like most people do ensuring quality work on time
Relevant Skills and Experience
CUDA, Finite Element Analysis, FPGA, Matlab and Mathematica, Verilog / VHDL
Proposed Milestones
$250 USD - .
Hi! My name is Fábio Pereira Benjovengo and I am a professor of electrical and computer engineering in Brasil since 2011.
I am using this app to rise my income and I have worked with principal component analysis in 2013-2015.
I am sure I can deliver what you need in a week.
Feel free to ask me any questions about it.
Regards,
Fábio Pereira Benjovengo, PhD.