What is needed:
My purpose is to use SVM(Support Vector Machines, a machine learning algorithm) to identify potential customers (who are the visitors) to any website. I am a lab teacher(with a degree in MS Computer Science), department faculty has assigned a task to make multiple,fresh and live environment programs that will be act as a lab manual to the BS Computer Science students. Problem described above is just one example among 19 others.
Right now, I got stuck in using MKL library. I am using python scikit library along side with MKL. I have successfully created multiple programs using only Scikit, but not able to counter MKL from the start.
What I am doing:
Right now, I am trying to set the intensities for 2D - points and then use 2D Gaussian distribution for those points, I am feeding this to the kernel(as Gaussian distribution gives good idea about probability of inputs) as input to my SVM. I am using leastsq function it is working well. But following is issue is coming.
Intel MKL ERROR: Parameter 6 was incorrect on entry to DGELSD.
Problem and Needs:
While searching for the solution, I find that, there is a little bug (that I can't figure out) with MKL. Please advice a solution, plus provide link to any library that is developed by Intel that make computations lighter in Machine Learning, as students have to run the programs on their home PCs.
Final Result Need:
At the end, as a live demonstration, this will run on a live server. I will access analytics of http://militarybases.co and perform the operations. At the end, students will know about a probability that specific visitor have a certain level of interest in the site (We have taken permission from militarybases administration).
Thanks for your time and help,
Jeana Eilts - Lab teacher at Unversity of Arizona, Tucson
PhD Scholar at University of Arizona, Tucson
Contact Information:
Address:1980 E. River Rd., st 100, Tucson, AZ 85718, United States
Phone: 520-202-4857
Personal Email: JeanaEilts1978@gmail.com