Early next year I will be taking the Coursera course "Computational Neuroscience. Rather than using Python it will use MatLab. MatLab is an expensive product. I have an old copy of it on the computer that died about a month ago. I bought MatLab because I got tired of running Octave in line mode on top of Cygwin. I didn't want to buy MatLab again just for this one course so I gave Octave another look.
Octave still has the same look and feel I remember but now there is a GUI wrapper for it. DomainMath IDE provides that wrapper. My few minutes with it so far looks promising even though it has a few quirks I haven't figured out yet. DomainMath IDE is a Java project so you will need to have Java running. I've running Java version 7 update 45. I installed Octave in its default directory then I put the DomainMathIDE_v0.1.5.7z file in the same directory and unzipped it using WinRAR. DomainMathIDE.jar comes up and seems to work just fine even though right now I'm getting the error: "'javaaddpath' undefined". I'll track this down sooner or later. While I installed the java library when I installed Octave it appears I don't have it registered quite right. I haven't been able to make DomainMath IDE the default editor for .m files or even right click on the files and edit them with DomainMath IDE. I can live with this for now.
WinRAR is worth the $29. It handles all the strange compression formats I've run across so far. Maybe 7-zip would work as well. It certainly would work on DomainMathIDE_v0.1.5.7z.
I bought Tutorial on Neural Systems Modeling so I could get up to speed before class begins. It turns out it is one of the suggested books for the class. The website has the source files for the "MATLAB BOXes". That's certainly easier if you don't need to type code to learn from it. There are plenty of exercises to get you head into the material. It doesn't cover cable theory so it looks like I'll be exploring point neuron models for now where spike rate is modeled by a number rather than spikes. I can live with this an learn. Later I can move to more realistic models when I'm investigating aspects this model doesn't cover. For now it's the Rosenblatt model, the Perceptron or a variation of it.
Sunday, November 10, 2013
Thursday, October 10, 2013
I may just be in a bad mood.
I don't know why people turn simple problems into hard problems.
About a week ago last Monday I fell asleep watching TV. When I woke up the power was out. The following morning when I went to turn on my computer it wouldn't turn on. This is my second computer on an APC UPC to die from what I guess was a power surge. OK, that's fine. Some things happen.
In 2012 I bought a developers version of SQL2012. It wouldn't install on my computer no matter how much I tried to get it installed. I think a trial version of 2008 had left something in the registry or on disk or somewhere so that no matter what I did 2012 wouldn't install. I bought a new faster computer, an i7-3770 processor with 8GB high speed memory. Unfortunately the video processor was slower than on my old machine so I didn't want to use it for gaming (5.1 subscore as opposed to 5.7 on my older machine.) After I did what I wanted to do I let the machine sit, collecting dust.
So when my computer died and I had time, a week ago this Saturday I moved my "new" machine over to my primary workstation and started bringing up to speed. The first problem I encountered was Trend Micro complaining about my virus scanning subscription being out of service and asking me if I wanted to renew. Of course I wanted to be protected so I went on the Best Buy site via the popup and renewed my subscription. After I did what I needed it asked me again. I thought it was
just more steps in completing my subscription but it was the subscription for a computer I gave my parents but they are no longer using. Even so, after saying I had renewed the software said I wasn't protected because my card was accepted.
I called Best Buy. I tried to explain the problem to them. I didn't want to pay for a one year extension to April 2015 for virus scan on a computer that won't be on network under my name ever again and I wanted the subscription I had paid for to be activated. After an hour or so of being shuffled around I ended up with the Geek squad. He said he could take care of the virus software but I'd need to talk with an account executive about getting the extra subscription removed. I had a meeting with my sister so after making sure everything was going to be handled I walked away while they had remote control of my computer.
When I returned I found out they had installed a trial version of the software. I reviewed my bank account and didn't find the payment so I waited until the next day, I was trying to get all of the software in place so I could do my "Computation investing" homework. OK, the next day my bank account had the cost of one subscription taken from it.
This evening I called Best Buy again after verifying the purchases on their website. I said I needed the serial number for Trend Micro because they had removed it from my computer when they replaced the software with a trial version. After some discussion they gave me the serial number for the subscription I purchased on Oct 6th. The serial number wouldn't work. After being bounced to the Geek squad again I was told the problem I didn't have the correct version of the software on my machine.
After they installed the correct version I noticed the expiration date is 9/27/2014. I asked why a one year subscription purchased on 10/6/2013 expired on 9/27/2014 and was told that was because my original subscription started on 9/30/2012. OK, my subscription hasn't been valid since 9/?/2013. If it wasn't valid then why was my 10/6/2013 subscription wasn't valid then why is it counted in the 1 year subscription? I don't get why a one year subscription staring on 9/30/2012 with one year extension would end on 9/27/2014. This is a bit like penny shaving for copper. Maybe one day a year isn't worth my time and effort but for them it means getting paid for 365 years subscription while only providing 364.
Yes, I'm stressed by issues at work where a promise was made yesterday so my customers would get supported after the upgrade of major software but now are not, a loss of, by their figures, about 80,000 hours of labor where this could be resolved by at most 10 hours of labor but they claim a major upgrade is a good time to move to best practices. My DBA supports them and may have instigated this turn of events. It's my employer's business. I just don't like all the noise.
I can't finish my Computational Investing class when I need to spend my evenings decompressing and preparing for another upgrade in another department where they've decided they didn't want to support the computer I had my employer buy a couple of years ago to support the volume and timeliness of data transfer from their system to ours. It turns out they have a 5 day month end freeze on the shared computer they are moving me to even though we have a 24 hour deliver SLA. Even so the data processing is in best case twice as long as current. It's too much noise.
About a week ago last Monday I fell asleep watching TV. When I woke up the power was out. The following morning when I went to turn on my computer it wouldn't turn on. This is my second computer on an APC UPC to die from what I guess was a power surge. OK, that's fine. Some things happen.
In 2012 I bought a developers version of SQL2012. It wouldn't install on my computer no matter how much I tried to get it installed. I think a trial version of 2008 had left something in the registry or on disk or somewhere so that no matter what I did 2012 wouldn't install. I bought a new faster computer, an i7-3770 processor with 8GB high speed memory. Unfortunately the video processor was slower than on my old machine so I didn't want to use it for gaming (5.1 subscore as opposed to 5.7 on my older machine.) After I did what I wanted to do I let the machine sit, collecting dust.
So when my computer died and I had time, a week ago this Saturday I moved my "new" machine over to my primary workstation and started bringing up to speed. The first problem I encountered was Trend Micro complaining about my virus scanning subscription being out of service and asking me if I wanted to renew. Of course I wanted to be protected so I went on the Best Buy site via the popup and renewed my subscription. After I did what I needed it asked me again. I thought it was
just more steps in completing my subscription but it was the subscription for a computer I gave my parents but they are no longer using. Even so, after saying I had renewed the software said I wasn't protected because my card was accepted.
I called Best Buy. I tried to explain the problem to them. I didn't want to pay for a one year extension to April 2015 for virus scan on a computer that won't be on network under my name ever again and I wanted the subscription I had paid for to be activated. After an hour or so of being shuffled around I ended up with the Geek squad. He said he could take care of the virus software but I'd need to talk with an account executive about getting the extra subscription removed. I had a meeting with my sister so after making sure everything was going to be handled I walked away while they had remote control of my computer.
When I returned I found out they had installed a trial version of the software. I reviewed my bank account and didn't find the payment so I waited until the next day, I was trying to get all of the software in place so I could do my "Computation investing" homework. OK, the next day my bank account had the cost of one subscription taken from it.
This evening I called Best Buy again after verifying the purchases on their website. I said I needed the serial number for Trend Micro because they had removed it from my computer when they replaced the software with a trial version. After some discussion they gave me the serial number for the subscription I purchased on Oct 6th. The serial number wouldn't work. After being bounced to the Geek squad again I was told the problem I didn't have the correct version of the software on my machine.
After they installed the correct version I noticed the expiration date is 9/27/2014. I asked why a one year subscription purchased on 10/6/2013 expired on 9/27/2014 and was told that was because my original subscription started on 9/30/2012. OK, my subscription hasn't been valid since 9/?/2013. If it wasn't valid then why was my 10/6/2013 subscription wasn't valid then why is it counted in the 1 year subscription? I don't get why a one year subscription staring on 9/30/2012 with one year extension would end on 9/27/2014. This is a bit like penny shaving for copper. Maybe one day a year isn't worth my time and effort but for them it means getting paid for 365 years subscription while only providing 364.
Yes, I'm stressed by issues at work where a promise was made yesterday so my customers would get supported after the upgrade of major software but now are not, a loss of, by their figures, about 80,000 hours of labor where this could be resolved by at most 10 hours of labor but they claim a major upgrade is a good time to move to best practices. My DBA supports them and may have instigated this turn of events. It's my employer's business. I just don't like all the noise.
I can't finish my Computational Investing class when I need to spend my evenings decompressing and preparing for another upgrade in another department where they've decided they didn't want to support the computer I had my employer buy a couple of years ago to support the volume and timeliness of data transfer from their system to ours. It turns out they have a 5 day month end freeze on the shared computer they are moving me to even though we have a 24 hour deliver SLA. Even so the data processing is in best case twice as long as current. It's too much noise.
Sunday, October 6, 2013
Getting new machine configured for "Computational Investing"
I'm under a bit more stress than I'd like. A lot depends upon me at work and there is little time. These minor annoyances don't help. I end up stalling when things don't go smoothly. None the less I learn quite a bit every time I have to deal with this stuff. This time around I learned I need to learn how python modules are normally installed.
I was lucky that WinRAR lets me install the software on a new machine using the same key. I'd have bought a new copy had I intended to use it on my old machine again. I don't want to take advantage of the company, especially when I like their product. It turns out the exe files mentioned below are auto-extract libraries. They can be explored with WinRAR. I'll probably look this a bit more when I get the time.
I thought I'd go ahead and upgrade the packages to the latest versions. It turns out I'm not up to speed yet so I have to rely upon the exe files. Several of the packages have new versions but I don't know how to install them. Rather than learn how to install and compile packages from zipped tar balls I used the most recent version existing as an exe file. My old machine was old enough that the dot net 1.1 dll was registered. My "new" machine is still old enough to be Win 7 but not so old that it had
msvcr71.dll registered. I found it in another directory and copied it to the setup directory I used to build the QSTK stack for the "Computational Investing" class. While installing pyparsing I got a couple of errors about not being able to register something. I played around a bit then pressed on. It didn't matter.
Here's the files I installed in the order I installed them:
I was lucky that WinRAR lets me install the software on a new machine using the same key. I'd have bought a new copy had I intended to use it on my old machine again. I don't want to take advantage of the company, especially when I like their product. It turns out the exe files mentioned below are auto-extract libraries. They can be explored with WinRAR. I'll probably look this a bit more when I get the time.
I thought I'd go ahead and upgrade the packages to the latest versions. It turns out I'm not up to speed yet so I have to rely upon the exe files. Several of the packages have new versions but I don't know how to install them. Rather than learn how to install and compile packages from zipped tar balls I used the most recent version existing as an exe file. My old machine was old enough that the dot net 1.1 dll was registered. My "new" machine is still old enough to be Win 7 but not so old that it had
msvcr71.dll registered. I found it in another directory and copied it to the setup directory I used to build the QSTK stack for the "Computational Investing" class. While installing pyparsing I got a couple of errors about not being able to register something. I played around a bit then pressed on. It didn't matter.
Here's the files I installed in the order I installed them:
Python 2.7.5
python-2.7.5.msi
numpy
numpy-1.7.1.win32-py2.7.exe
scipy
scipy-0.12.0-win32-superpack-python2.7.exe
dateutil
python-dateutil-1.5.win32.exe
pyparsing
pyparsing-2.0.1.win32-py2.7.exe
matplotlib
matplotlib-1.3.0.win32-py2.7.exe
pandas
pandas-0.12.0.win32-py2.7.exe
setuptools
setuptools-0.6c11.win32-py2.7.exe
cvxopt
cvxopt-1.1.4.win32-py2.7.exe
scikit-learn
scikit-learn-0.14.1.win32-py2.7.exe
scikit-statsmodel
statsmodels-0.5.0.win32-py2.7.exe
qstk
QSTK-0.2.6.win32.exe
Saturday, October 5, 2013
Computer dies, no shared backup
Well, my computer died last Monday evening. I was delaying HW3 on "Computational Investing, Part 1". I had completed the framework and was ready to begin testing. I went down to watch some TV. When I woke up the power was off. The next morning I tried to turn on my computer but it wouldn't turn on. I didn't have time to work on setting up my secondary computer. I hadn't turned it on for about a year. I began setting it up this morning but wanted my virus software enabled before I did much. I followed the instruction and it lead me to best buy where I paid for the subscription then it lead me to a second page and it looked like another step in the process but it was for an extension of the subscription for my parents but they don't have a computer any more. I called best buy to get this handled and after getting bounced around and a couple of hours it looked like it was going to get handled. I had to leave. When I returned Trend Micro had installed a trial version. So here I am having just paid for the subscription renewal and for a second subscription renewal for a copy I didn't want (good through April 2015) and I have a trial version good for 30 days on my machine.
Well, I've got to reinstall a bunch of software and get my "Comp Investing" homework done so I'm not going to be very happy when I get to work on Monday and have to deal with more computer problems. Ah well. I'm glad I documented the installation of the "Comp Investing" software on my blog. I'll go back and find it and update to the latest versions as I do. When I do I'll provide an updated version. At least I have SQL Server installed on this machine where I couldn't get it installed on my other machine. This machine is faster but has slower graphics. I may have to do something about that.
Well, I've got to reinstall a bunch of software and get my "Comp Investing" homework done so I'm not going to be very happy when I get to work on Monday and have to deal with more computer problems. Ah well. I'm glad I documented the installation of the "Comp Investing" software on my blog. I'll go back and find it and update to the latest versions as I do. When I do I'll provide an updated version. At least I have SQL Server installed on this machine where I couldn't get it installed on my other machine. This machine is faster but has slower graphics. I may have to do something about that.
Monday, September 2, 2013
Week 2 of Computational Investing Part 1.
I guess I did the hard part of the week 2 homework by installing the software on Saturday. Again, much of the territory isn't new to me. It was nice to review since I'm going to put some of this into practice, or so I hope.
Today I learned a bit more about the QSTK software and the Matplotlib tool. I just copied and pasted the text from The Quant Soft Took Kit Tutorial and could rapidly rebuild the graphs in the demo. Normalizing returns and producing scatter plots are very easy. Since the data is retrieved from Yahoo Finance I'll not give it here. Review the tool if interested. Be polite and save data locally after retrieving it once if you intend to work with it often.
Normalized returns look at the price and change relative to some point in time, that is Price(t)/Price(i) where i is a constant. Most of the time the price is normalized to i=0 and t ranges from 0 to some larger value but one can normalize to a time in the middle of the graph if one likes.
A scatter graph plots the change of two items relative to each other based upon some interval. It shows how correlated the two items are and whether the are positively or negatively correlated. The greater the deviation from the norm the less correlated they are.
Since I'm learning, I won't put in my own thoughts. It's hard enough to suppress them while learning. Once I have a good grasp of the subject matter I'll put my own stamp on it.
Even if you aren't interested in the stock market the NumPy and SciPy pages as well as Matplotlib if you are interested in doing scientific computation and plotting. I wish the "Computing the Matrix" instructor had used Matplotlib rather than his home grown plotting tool. Matplotlib can show a plot in it own window or produce a PDF. It may be able to produce files in other formats, I haven't taken the time to look.
Today I learned a bit more about the QSTK software and the Matplotlib tool. I just copied and pasted the text from The Quant Soft Took Kit Tutorial and could rapidly rebuild the graphs in the demo. Normalizing returns and producing scatter plots are very easy. Since the data is retrieved from Yahoo Finance I'll not give it here. Review the tool if interested. Be polite and save data locally after retrieving it once if you intend to work with it often.
Normalized returns look at the price and change relative to some point in time, that is Price(t)/Price(i) where i is a constant. Most of the time the price is normalized to i=0 and t ranges from 0 to some larger value but one can normalize to a time in the middle of the graph if one likes.
A scatter graph plots the change of two items relative to each other based upon some interval. It shows how correlated the two items are and whether the are positively or negatively correlated. The greater the deviation from the norm the less correlated they are.
Since I'm learning, I won't put in my own thoughts. It's hard enough to suppress them while learning. Once I have a good grasp of the subject matter I'll put my own stamp on it.
Even if you aren't interested in the stock market the NumPy and SciPy pages as well as Matplotlib if you are interested in doing scientific computation and plotting. I wish the "Computing the Matrix" instructor had used Matplotlib rather than his home grown plotting tool. Matplotlib can show a plot in it own window or produce a PDF. It may be able to produce files in other formats, I haven't taken the time to look.
Subscribe to:
Posts (Atom)