stdbrouw

73 ms·

751.
by stdbrouw 12y ago
> If an professional industry can be grown around child care, I think its better than having parents do infinite amount of harm to their kids. We know that if you drink lots of alcohol while pregnant, you're going to harm your child
752.
by stdbrouw 12y ago
Also see http://www.ledger-cli.org/ I like how Ledger is essentially just a system for doing arbitrary calculations on "buckets of quantities of stuff" a.k.a. accounts with various currencies and assets. I'm
753.
by stdbrouw 12y ago
Whether using animals is necessary for the research is one thing. I'd also quite like to know whether the research is necessary to start with. I assume that the research design when animals are involved is generally of a higher quality
754.
by stdbrouw 12y ago
So do you feel the pitfalls of probabilistic programming / Bayesian data analysis are potentially more severe than those of traditional methods? I don't get to use that much Bayesian analysis in my work, but I've been toying
755.
by stdbrouw 12y ago
Well, MCMC is pretty close to brute force so it's never going to be very fast. But the diagnostic tools are really quite good these days – tools like Stan and PyMC make it easy to see diagnostic plots that allow you to check for conver
756.
by stdbrouw 12y ago
Nitpick: while that paper does mention the basic idea that there's no significant difference between 0.051 and 0.049, it's more about another common mistake: testing the difference between two groups by performing two separate sta
757.
by stdbrouw 12y ago
In many languages, including French, sometimes the past and present tense sound the same and are even spelled the same. The world hasn't exploded yet. People figure out ways to get their meaning across. Also, dictionaries based on acce
758.
by stdbrouw 12y ago
Isn't this just the linguistic equivalent of micro-optimization, though? There's tons of grammatical (e.g. differences in the subjunctive between verbs that use the auxiliary etre or avoir) and spelling oddities (le hanche but l&#
759.
by stdbrouw 12y ago
The referenced graphs measure performance and perceived ability on 4 different tasks. You're right that ideally you'd pool this data to get to an overall correlation, but for the point the author makes, eyeballing it and taking a
760.
by stdbrouw 12y ago
The more common advice is to get up at the same time every morning, regardless of when you went to bed.
761.
by stdbrouw 12y ago
Why does every data science infrastructure provider want to be full stack? I really don't want to use an online IDE, I really don't want to be locked into one particular way of distributing workloads, I really don't want to m
762.
by stdbrouw 12y ago
Even if we ignore the issue of residual confounding for a moment, how is one IQ point "damning"?
763.
by stdbrouw 12y ago
> Turns out, this was good preparation for real life in the competitive fields that many of us chose. I wonder if this is self-reinforcing. When you're used to huge amounts of stress from school, it seems natural that companies will
764.
by stdbrouw 12y ago
We should be equally worried about being overly dependent on clinical methods (human judgment) for important decision making. When looking at an algorithm, you can at least see how it works. Not so easy for "trust me, I'm a doctor
765.
by stdbrouw 12y ago
There might be something there, but I don't think cuts are a very good example. For cuts, the advice is always: close 'em up if they're big, leave 'em alone if they're small. Ditto for allergic reactions: many diffe
766.
by stdbrouw 12y ago
> I really wish we could replace our 12 years of mathematical education, which I think is overall quite useless[2], with 12 years of linguistic education I wonder why it is that people can never just say "you, know, X is quite impor
767.
by stdbrouw 12y ago
> Having to rephrase something because someone already said it that way is an insane waste of time from a bygone era. Or, a great way to learn and internalize a concept rather than just regurgitating it.
768.
by stdbrouw 12y ago
The existence of one plausible explanation (good model fit) does not automatically invalidate another plausible explanation (mechanism of action).
769.
by stdbrouw 12y ago
It's really not. Statisticians usually receive only minimal training in statistical programming, data cleaning, data gathering and warehousing and so on – yet those are all essential to data science. On top of that, when dealing with b
770.
by stdbrouw 12y ago
In the wild I think you'd be more likely to use something like PyMC or Stan, which work like Prolly but support arbitrarily complex models. Still, love the idea. To learn more about probability theory and Bayes, "Probability Demys
771.
by stdbrouw 12y ago
Containers don't solve my particular problem, hence containers suck. What wonderful reasoning.
772.
by stdbrouw 12y ago
There are absolutely ways to deal with what you call interconnected data, as I mentioned earlier: paired tests, corrections for autocorrelation, nonparametric and bootstrap methods for non-normal data and so on. But barring any examples of
773.
by stdbrouw 12y ago
Hmmm, again, there are certainly ways that interactions between visitors can cause statistical dependence, but not in the specific case you mention. Let's take an A/B test on a referral funnel. If a user invites all of his friend
774.
by stdbrouw 12y ago
Could you expand on what you mean by "interconnectedness within website traffic"? If one person visiting the site has no influence on other people visiting the site, then measurements of their behavior will be independent. If Face
775.
by stdbrouw 12y ago
The problem is that p-values are begging to be misunderstood, and in fact you cannot use them as a decisionmaking procedure without "misinterpreting" them – after all, you're deciding whether to accept the hypothesis P(HA|D)
776.
by stdbrouw 12y ago
Banning p-values makes sense to me, as they force you to declare an effect as either significant or not significant, rather than looking at the preponderance of the evidence and building knowledge over multiple experiments. It also leads us
777.
by stdbrouw 12y ago
"The probability of this parameter being in the interval" is what's called a credible interval, and it will coincide with the confidence interval if you assume a uniform prior, that is, if before running the experiment you
778.
by stdbrouw 12y ago
It's an improvement, but not nearly as easy to use as leafing through a physical book. I'm thinking of the use-case where you're solving exercises in one place but also want to sometimes go back to the theory or to a referenc
779.
by stdbrouw 12y ago
Hm, I'm going to guess that the Kindle is probably the most popular e-reader, and it has the ability to add notes, highlights and bookmarks, both with the hardware device and in the apps for tablets and computers. It also doesn't
780.
by stdbrouw 12y ago
The criticism of sampling in this blogpost is way overblown. (1) there's virtually no difference between the accuracy of a sample size of 10k and one of 100k, and those are the sample sizes you're usually working with (2) in parti

More ›