darkxanthos

I don't see this being the same market at all. Gopro is marketed at adventure and sports. Polaroid is pushing this toward the average Joe. With a $99 price point it's well positioned to displace gopro.

It's important to note that the author in question is one of the largest forces pushing Bayesian Statistics forward. I'm not arguing your points but just pointing out that it's arguably against his self interests to put something like this out. Again, just adding the appropriate shading for those not in the know.

Interesting. Thanks for that

Does it seem weird to anyone else that the first example is approximating XOR?

Just to point out, if you really will have a tiny list and that's knowable, it's possible this example would be best with a straight linear time check. It could be fewer operations than hashing a string and looking it up. Practically pedantry though.

This could just mean most people suck at practice. There appear to be very specific ways to practice different skills. Practicing wrong means engraining bad habits and effectively creating an artificial plateau.

This exactly. This article didn't even try to objectively validate the report. It's also provably false in other areas like already not very busy roads in the high desert in SoCal. Add a new road to Apple Valley and I bet it gets used but very little there's just no demand.

This report was highly biased since I'm sure no one built a road where people didn't need one and looked if demand were the same. Instead this is only about building roads where they were deemed to be critical enough to warrant a ton of spending and where the built out supply probably lags demand.

Yep. I think what he's really trying to do is find a definition of morality that's useful, not necessarily complete. As models go, it was useful enough to change my thinking a bit.

well the range is really $36 - $1,041 per month. That seems reasonable depending upon what you need.

A lot are. I used to want to just because it was cool. Now that I actually do though my main use case is sifting through hundreds of GB of unstructured data. I use Hadoop to get the data into a structured form that I can then load into Redshift. It's awesome knowing that just about anything I throw at Redshift, it can handle.

Thanks for this comment. Just to be clear I am not saying it isn't necessary but am genuinely curious what I'm missing. I think a big influence on my comment is using ideas that touch on linear algebra and possibly even doing the computations one might end up doing via linear algebra but without knowing it.

For example- Least squares regression. Totally use this. Even took a semester in college on just regression. The linear algebra underpinnings though haven't never been shown except for a quick blurb in my linear algebra text book. I still understand the concepts of fitting a model and when it's a bad fit (such as non-normal distribution of residuals, co-linearity) but the theoretical underpinnings are more fuzzy to me.

Representing features as vectors, sure. But that's also a pretty superficial use of linear algebra since from that point forward I'm using something on the trig side to compute results (at least in clustering).

I also tend to lean rather heavily on probability and bayesian approaches to many areas. So Naive Bayes classification is a love of mine, finding ideal parameter values given data coming in becomes an online updating multi-armed bandit problem to me (which also doesn't require explicit linear algebra). A lot of my work is also in experiment design and analysis and for this I use a mixture of Bayseian and frequentist statistical testing.

Canned packages out of the box with parameters to tweak that I can cross validate to evaluate how well my model is working. If I happen to venture out to other models I'm probably reading up on common pitfalls and how to test for them.

To me, it's entirely possible that the gap between you and I is due to experience and even just differences in training/learning (including but not limited to the quantity of it). These discussions are important for me since they help to inform my future learning aspirations.

Thanks for the response. That gives me something new to learn/look into.

What are people using Linear Algebra for in Data Science? Aside from the stock representing words as N-dimensional vectors I mean?

I ask because I do this kind of work as my J.O.B. and every skill he refers to I totally see the necessity of except this one.

I don't think we can say it's clearly working... This seems like an extraordinary claim and it's hard to know if there's any truth fudging.

If you have the skills, build your own. It's dirt cheap, again, given the right skills.

Ive heard of a similar story second by my ex wife about her cousin when he was three. The short version is they both walked into a liquor store one day and her cousin recognized a woman as being from a construction site where she worked. He started discussing details with her that a three old just wouldn't know. After the fact my ex would ask him questions about his "past life" after enough prodding he'd become visibly uncomfortable and sad about talking about his death. Once he turned five or so he stopped remembering.

I've always wanted to believe, and given its second hand nature and the credibility of the source I count it as weak evidence at best. Obviously nothing even close to proof, but I've always found these stories intriguing since then.

You're right. My bad. Thanks for correcting me.

Well a homomorphism applies to a group and a group can only have one operator by definition. That won't change probably. Rings have two operators, fields are rings that support cancellation and integral fields I believe support inverses as a rule and where you'd get "everything" I believe.

The problem isn't so much centered on the "shoulds"/ideals it's centered on reality and a more game theoretic approach. Instead of everyone crossing their heart to be good... how do you make the desired outcome have the least friction in being realized?

I'm currently pursuing my math degree here (1.5 years) and have enjoyed it and learned a ton. Currently enrolled in Abstract Algebra which is already a tough course but then also add t ok it that the online classes are more intense 8 week "semesters" and it equals pain. But it's rewarding pain. This seems like a logical fit for them and I think it could be a great experience for those who are interested.

This is timely for me as I've just started a course in Abstract Algebra after taking a class on proofs. There doesn't seem to be a great way to get through this course without full understanding of the concepts since it is so proof laden.

The idea struck me yesterday that there's no real reason one couldn't teach this in elementary school or Jr. High and maybe that would be amazing for students? These kind of courses shape the way you reason.

Not at all... He states merely that if someone has the option to have a control and chooses not to, you should be concerned. Specifically he hits on the fact that some fields can't have a control (sometimes just due to ethical reasons).

Yup. When you have a small amount of traffic this logic really just points out your primary goal should be either getting more traffic or only designing tests that should have a large effect.

Some people reply to this by saying "Well then we'll just ship it and measure results without testing." That is a test... but it's poorly designed. If you can't see the improvement in a controlled experiment, it's rare you'd be able to see noticeable improvement in before/after.

Am I the only one astounded that we're still having this conversation? The fact that it's still around shows me it has indeed been "the answer" for some. The fact that it hasn't become ubiquitous is plenty of proof that it also is not a one size fit all solution, as if any solution is.

He even says this is possible via Bayesian experiment design:

FTA: Bayesian experiment design: With Bayesian experiment design you can stop your experiment at any time and make perfectly valid inferences.

You're right. The one exception though is with Bayesian statistics you can estimate an effect size using your experiment results using a credibility interval.

If the differences are drastic enough you can still get value from split testing. Incremental changes are just probably not going to bring you much luck.

I use simulations so I can avoid math and analyze the power of any distribution I can model in a consistent manner.

I wrote something about it here: http://www.databozo.com/2013/10/12/Finding_a_sample_size_for...

EDIT: typo

I do this professionally as my sole job. This is one of the very few papers I've read that seem completely legit to me. I especially love their point on necessary sample sizes to get to a 90% power.

Yup. I personally love Ruby but as my skills have specialized more towards data science spectrum I've found Python to be the place where all the libs are.

Ruby and Python are also similar enough that while there are a ton of differences they won't be difficult to digest.

Focus less on: * Comparing them to you * Changing them

Focus more on: * Identifying and leveraging their strengths * Position your self as a Silvers type leader. Be a first follower to any of their ideas that you think can be tweaked to hold water.

That last one may be non-obvious. Learning to be a first follower promotes humility in yourself, supporting others, and showing others how to lead and follow. When others lead their 'laziness' (I hate that word) begins to dissipitate.