3/18/2008 3:17:00 AM
Continuing on in the effort to add a Bayesian filter to our "analysis" program, I show you the basic steps of configuring our filter so we can train it to sift through past performance data. It's pretty straight forward, but Bayesian filters pretty much are that way. There is nothing complicated about them, but the results will seem almost like magic. I show at the end of the screencast the training app used to train our filter and will provide a bit more detail on that part of the solution in the next screencast. That is also when I will show you how we actually use the filter to analyze data and filter the good from the bad.
You can download the source here.
Note: I accidentally set the screencast to start with the volume muted. So you will need to click the mute button to get the sound working. Sorry - I will try and upload a good player config later. |
Double-click for full screen |
Subscribe