pps43

23 ms·

211.
by pps43 8y ago
There are many different credit scores. The range of valid FICO scores does not include zero, but internal acquisition and behavior scores that banks use might. But in this case it's irrelevant because average balance is independent va
212.
by pps43 8y ago
> I don't see the issue with those examples Let me elaborate then. Average credit card balance predicts probability of default. Joe who maxed out his credit card is higher risk than Jane who pays back her entire credit card balance
213.
by pps43 8y ago
It's not only NASA. Zero is just not a reasonable default value for an average. Imagine average credit card balance for FICO score, average blood pressure, average temperature in a freezer, etc. UB is not the only way. You can have NUL
214.
by pps43 8y ago
Antisemitism in Soviet Union followed from the way they understood equality. For example, they look at the distribution of university students and notice that Jews are over-represented while, say, Uzbeks are under-represented. That's n
215.
by pps43 8y ago
> `average = 0` with no elements Imagine a Mars lander program that looks at the average of altitude measurements to decide when it's ok to jettison the parachute, sees no measurements, and decides that altitude is zero.
216.
by pps43 8y ago
I used wxMaxima. Worked fine for what I needed it for (integration, solving systems of equations, etc.) In some cases it even solved problems that Mathematica couldn't.
217.
by pps43 8y ago
> obscurity is net-negative for systems that need to defend against strong attackers If that's the case, why doesn't the NSA publish Suite A algorithms?
218.
by pps43 8y ago
Good security should not depend on obscurity, but it does not mean that security through obscurity never works. It's still better than complete transparency.
219.
by pps43 8y ago
YCombinator can go out of business and bring HN down, not all internet memes survive long, your disk can be lost in a fire, etc. I'm not saying blockchain is better, just that it's at least reasonable. Already distributed and can
220.
by pps43 8y ago
Now you have to think about backup, transporting across borders, etc.
221.
by pps43 8y ago
That's exactly what the blockchain is doing, distributing the data across multiple devices.
222.
by pps43 8y ago
And store the encrypted passwords where?
223.
by pps43 8y ago
On reflection, this is probably not as stupid as it sounds. The point is using blockchain to store your (encrypted) data. It's not feasible to backup everything this way, but a list of passwords is short enough. If all you do is open s
224.
by pps43 8y ago
> it's considered a bad thing to have problems that are essentially unsolvable for most of the class. It's really sad, for two reasons. First, most problems in most AoPS books are solvable (or at least should be) for most of t
225.
by pps43 8y ago
I bought their books and got my kids to go through them, in order, solving problems along the way. This proved to be very effective, both in my subjective opinion and based on the results of the standardized tests. Those books are head and
226.
by pps43 8y ago
All banks already use behavioral scores in credit card line management. Mortgages are a different story because there's little they can do once after the underwriting. That said, independent variables that go into behavior risk scores
227.
by pps43 8y ago
The baseline (FICO) has AUC of 68.3% which looks low. This may be because the analysis is performed not on the entire through-the-door population, but only on the customers that passed the creditworthiness check (which is using FICO). In su
228.
by pps43 8y ago
That game is common in credit scoring classes for fresh analysts. The class is split into small teams. Each team is given ten anonymized, but real, credit applications and corresponding credit bureau pulls. Five from customers who subsequen
229.
by pps43 8y ago
Depends on the location. In Japan it's a waste of time and resources, in US it can save lives. I was field stripping and re-assembling AKM in school (just like everyone else in the class), on the clock, under watchful eye of a real war
230.
by pps43 8y ago
People like myself with both presbiopia and myopia can still comfortably view something close to the nose (without glasses).
231.
by pps43 8y ago
Nearsightedness does not go away after 40 (presbiopia does not cancel out myopia).
232.
by pps43 8y ago
Very difficult. It's much easier to train another model. Those things are mostly black boxes. Even the developers might not understand completely how they work.
233.
by pps43 8y ago
FPTP satisfies two out of three Arrow's conditions (unanimity and universality). It does not satisfy independence of irrelevant alternatives, but in this sense only dictatorship is better (which satisfies all three). If you exclude dic
234.
by pps43 8y ago
I don't think first past the post is as bad as the article paints it. Every voting system is suboptimal in some sense due to Arrow's impossibility theorem. With FPTP you're likely to end up with two candidates who have positi
235.
by pps43 8y ago
Some of those are simply wrong. For example, the section on power law says "The central limit theorem does not apply and there is thus no 'average' earthquake. This is true of all power-law distributions." But power laws
236.
by pps43 8y ago
It's hard to argue that "bear arms" means something other than "carry a gun". I don't think open vs. concealed matters in this context. It's probably even easier to "play a cop" with open carry.
237.
by pps43 8y ago
I think that instead of adding one marble at each step, it's better to add a fixed percentage of existing marbles, e.g,, 0.25%. This will better approximate the processes that are being modeled. But then the effect shown in the article
238.
by pps43 8y ago
I'm talking about pseudocounts (add one success and one failure) [1] or maybe Agresti-Coull estimator (add two successes and two failures) [2]. [1] https://en.wikipedia.org/wiki/Rule_of_succession [2] http:/
239.
by pps43 8y ago
Wouldn't it be easier to just add one success and one failure and calculate (events+1)/(trials+2) instead of events/trials?
240.
by pps43 8y ago
But you know neither the chance of the new rocket being explosion proof, nor the chance of the rocket blowing up half the time. You built the rocket, launched it once, and it did not blow up. What should be integrated to get the probability

More ›