Week Ten:

Besides the problem of the time factor closing in around me...I had relatively few problems. I struggled for a while about how to make all of the data files work, and I think that I came up with the right solution.

I've been writing readMe files...nothing too extensive, just little files of what everything is and how it all works. But it's a little bit of a problem to decide what all goes in there! That's all though. Few problems for the last week.

Week Nine:

So I guess I was a little embarrassed to admit this last week. I was very frantically trying to solve the problem of clustering in different numbers and types of dimensions at the same time. I was absolutely positive that I was missing something completely obvious, and I was just blind to whatever that something was. Got really frustrated, yelled at the computer and the project a couple times. I talked with Marie on Monday, and apparently that's a problem that hasn't been solved yet! Good to know that I'm in good company.

Also, this week I was dealing with the apparent problem that my spring forces were defective. I'll admit straight up, I have never taken physics of any kind. High School, College, none. I had no idea what the algorithms included in Prefuse were doing. While talking with James, who is going to continue the project when we're done, he gave me a few ideas to trace the error. Well, I did, and it turns out it wasn't an error! The problem had to do with the masses of the nodes themselves. It's mostly fixed, but it's still kind of jittery.

And let me reiterate the fact that I will die happy if I never see another abstract action again.

Week Eight:

Besides my enormous problems with, you know, stuff like balance, coordination and general not injuring myself, I've had a few problems.

I know I had problems, but I honestly cannot remember what they were.

Week Seven:

Oh Weka. What on earth are your class labels for? I must have them to cluster, but all they do is tell you how many clusters to have, even if I use the set cluster number method. It doesn't matter which Instance(s) have what class label, as long as they're there. That was the basis of my problems this week.

First I wasn't getting a run through of the clustering method. Then I was only getting one cluster. Then I figured it out. Oh, Weka.

Week Six:

All of my big problems for this week seem to be unrelated to the project, and mostly with residential life. But, I still had my moments in front of the computer! Creating all of the vectors for the reading in of the xml files was a chore...and so many loops. If I didn't have any experience with Iterators before, I sure do now. I realize that this is a great experience, but a small part of me wishes that I could have had a little bit more experience before this! But the best way to gain experience is to jump in, right? And I guess now I know that I can figure my way through a problem, and a little confidence is good.

So it appears that this "problem" section for this week is mostly musings, rather than distinct problems. I'm sure I'll have more to report next week. You'll notice that I've started my archive, and my Bryn Mawr site now mirrors this one, although it might get updated a little less often.

Problems for First Half