Hacker News new | past | comments | ask | show | jobs | submit login
2009 UCSD Data Mining Contest (ucsd.edu)
27 points by TriinT on May 24, 2009 | hide | past | favorite | 8 comments



I find it humorous that it is sponsored by FICO, the credit card industry friend. College students are those most approached by credit cards. They are also the least likely to understand how to use them appropriately.

FICO is really a system designed to train people how to live beyond their means, through the incurrence of debt.

The best way to avoid the shackles of debt is to have a really bad FICO score.


I disagree. That is irresponsible advice. The best way to avoid the shackles of debt is to be financially responsible. A good credit score is very important for more than just a good interest rate on your credit card.

Your FICO score affects what you pay for insurance (or even if you CAN get insurance).

http://www.iii.org/media/hottopics/insurance/creditscoring/

Some debt can be worthwhile, especially when you balance your opportunity costs. Renting versus owning: Save for 15-20 years (or more) before you buy a house outright or save for 2-5 years and use credit for the rest? Need a small business loan for growth? I'm pretty sure your personal credit will affect that too.

And if you think you will ever need a security clearance, you better have a good credit score.


Responsibility with your finances is the best bet, so I will qualify what I said with, "For irresponsible people, it's best to have a bad FICO score and not incur debt. Better to live a cash life." This often includes college students.

Also, insurance is for poor people. Rich people don't have insurance. For example, in Texas, you don't have to have car insurance if you can verify that you have $100,000 in cash available to compensate a victim of your wreckless driving. Debt is for poor people. Debt keeps people poor.

If you want to be rich, take on no debt, spend only cash, and earn more of it.


Actually, many rich people got rich by leveraging debt, not shunning it.


Yay! And I was looking for interesting opportunities to try out data mining tutorials recently discussed here in HN. http://www.autonlab.org/tutorials/

Now we have ready to analyze dataset, ability to get immediate feedback on performance of my model, and compare it with the rest of the world.

P.S. I'm new here.


Thanks for joining. Stay a while and contribute as much as you can. I'm a semi-newbie myself and I've learned a ton just by checking in everyday.


Can anyone decode this jargon: "The performance metric is the lift at 20% review rate."

By the way, ... the contest is only open to students.


From their page [http://mill.ucsd.edu/index.php?page=Datasets&subpage=Tas...]:

Scoring Predictions

Evaluation Metric: The evaluation metric for the E-commerce Transaction Anomaly Classification (Easy) is lift at 20%. If S is the 20% of examples that you think are most likely to have a positive label, then the lift at 20% is proportional to the number of positive examples in S.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: