470 karma · joined September 12, 2018
Hi.
Just a statistician/programmer.
You have a good day now.
There are two classes of sunscreen, chemical and mineral/physical.
The mineral/physical ones are group by two mineral zinc oxide and titanium dioxide.
You get a ghost cast using mineral but you can offset it by using a tan/colored version.
Mineral sunscreen are better because it stop UVB and UVA. It also doesn't require a waiting period after applying. Zinc provide wider spectrum of protection than titanium.
Clinical trials have 3 phases and it all have to be explained via statistic (design of experiment) and reproducible. They have standards for everything to enter the US drug market.
Lots of biostatistics stuff too.
Even when the drug enter the market, pharma companies stop testing the drugs. It is up to the FDA to keep track of side effects and toxicity that were not caught through the phases. (example: ambien sleeping drug)
The people at FDA NCTR are wonderful, highly recommend interning there for a summer.
They have some ML stuff. I've seen a project where they attempt to figure out what bug is in the food. One of the bug was midly toxic and they wanted a image classifier to figure it out via picture of a bug part. You can only have a certain ratio of bug parts in food.
On top of that, the Covid19 happened.
Some theorized that SA and Russian pumping more oil was a plan to kill USA's fracking and shale industry. I'm just hoping it'll force us to move toward renewable so we won't be dependent on these countries.
The name itself states this. Autocorrelation and autoregression. It regress on it's past values hence the "auto". We're interested on past value so each other value is dependent on the past.
An example of this is taking blood pressure. You would assume that taking blood pressure two consecutive days means that the previous day is highly correlated to the present day test.
Where as if you compare your blood test a year ago compare to today blood test it won't be as correlated. This is why ARIMA is dealing with correlation with time lags/lengths.
What you're doing is looking at autocorrelation base on time lags. That's what ACF and PACF graphs is displaying.
There are multivariate models but I don't know much about those. Most of the good resources are in the econometric domain. Multivariate time series within econometric, from what I've seen, is portfolio balancing.
For a general overview for statistic domain I would recommend:
For ARIMA I love this book:
Time Series Econometric by Levendis
For GARCH: Financial Modeling Under Non-Gaussian Distributions
If you want to learn more within statistic and time series in medical data: there is (1) longitudinal and (2) survival analysis. There are non linear time series but those are rare because most of our tools work within linear. There are also circular time series and temporal spatial statistic but I don't have any relevant knowledge in those to give you. I'm sure there are other that I don't know about within statistic.
Another interesting one is change point statistic https://en.wikipedia.org/wiki/Change_detection.
There is also a coursera course in time series that I've taken. I will post it here when I get off of work and better internet connection.
If you want an idea what forecast models out there you should read the papers from https://en.wikipedia.org/wiki/Makridakis_Competitions
There are 4 papers now and most of them are on statistical models which traditional dominating this domain. Datascience/ML models are slowing getting in there. M4 the best model was a highly tailor hybrid between ML/Stat technique the person who created it was employed by Uber and wrote an article about it.
The 5th competition m5 is currently underway and split into 2 contest. I'm eagerly waiting to read the paper on the results.
Intel tend to give turbo performance which can only run for so long before heat becomes a problem.
Do you have the data to back that up or is just a question that you don't have an answer to?
I think the statement brush off the high quality R packages that already in the ecosystem that is not in anywhere else. The author of element of statistic learning created glmnet package which took python years to even have ported. There are many other packages out there that Python does not have.
I am not going to argue it's the prettiest language.
But there are tons of packages and many found no where else in other language ecosystem.
If you're going to say just use rest/rpc then it just defeat the purpose of your initial argument.
And just look at springer or, gosh the other publisher escaped my mind right now, but these publishers have book on statistical subjects with R packages accompany the book.
It may be ugly but the packages are maintain by expert in the field of statistic and they may not be programmer. But they do dog food their package and use dataset to see the results. Likewise just read up on the Ranger Rpackage paper (https://arxiv.org/pdf/1508.04409.pdf). They test their output with the other randomforest package.
And it's silly to point this argument at just R when the same happen to Python. The bootstrap function in SciKit-Learn for the longest time didn't even really do bootstrap. The linear regression function automatically does shrinkage with no option to turn off.
No language is perfect. But I believe R have a place and especially in statistic. Many many wonderful expert statistician are maintain and creating R packages (eg Dr. Frank E. Harrell Jr. ) Many R packages have accompanying paper publish here https://www.jstatsoft.org/index
Erlang does it too.
If you want to have state you do recursion and call itself with the updated value.
fn me(iter) -> me(iter+1) end
Thank you Chris and Jose and all those who were and are involve with Phoenix.
Can you further explain this part? It sounds so out there with undersea cable.
The only thing I can think of is that wildfire cause erosion in soil which wash off into the sea and damage the end point?
Yeah... about that.
From a western news outlet point of view, it seems like Sweden is doing it wrong.
Covid19 does not need to be politicalize. We're in troubled times I don't believe we need people going against experts (science/doctors/etc...).
It's like choosing N.Korea as an example and asking why you should trust them with your civil liberties.
https://privacyinternational.org/sites/default/files/2017-12...
My previous company store coupon code on it (referrals). I prefer a hash nosql for that though (cassandra) but for a one man shop it's overkill.
---
edit:
IIRC postgresql have their own document storage like MongoDB. I just never had to use such thing because I guess I'm too OCD about how my data suppose to look like.
This is just one of the two competitions for m5. The other one is uncertainty.
Deep learning does not do this.
Data with less noises are what most deep learning and non statistical models does well. Meaning that image, nlp, etc.. deep learning does well. But data with lots noises/uncertainty/variance or even data that isn't large enough, such as time series, currently statistical models are still king (https://en.wikipedia.org/wiki/Makridakis_Competitions).
Even with healthcare you're answering a question/ hypothesis. This is where statistical models strength lies because all statistical models are hypothesis tests and vice versa. There are very little opportunity in healthcare where you would use deep learning compare to statistic. I've seen NLP can be of use but the majority of work in healthcare are inference/casuality base (this is why they use propensity model so much). I'm in this space public healthcare.
Historian see WWI as the war of chemists.
And WWII as the war of physicists.
He really tweet fight over other statisticians about the concept of black swan when it's in the statistic. I can understand if he wants more emphasis on it but he makes it as if statistic does not take this into account.
Black swan is something that is never seen. Seeing how there were many similar viruses coming out of China and one in the middle east already means this is not a unseen or will blind side.
Yes, unfortunately, it have to be political. There is no escaping it, when it comes to healthcare.
It is much more urgent when country keeps on creating new viruses and we become more globally connected.
They usually have to go through it again, FDA clinical trials (phase i, ii, iii) with different dosage to see what's effective. Increase in dosage can increase chances of side effect or create new ones.
There may be laws to get around or fast track that but pharma company have a dedicated department to figure that out. Just like that one company that causes opioid epidemic.
But it's also inescapable to ignore Erlang when you dive deeper into the actor model. Several good actor model books are Erlang only. Unfortunately many people, including myself, ever actually do anything hardcore with Actor Model. There is a good blog post about several skill stages of erlang/elxir and OTP was at the very top in term of few people actually uses it. I can't recall it the author =/.
Actually, I like Elixir package management system and command line tool is much better. They recently have the release command line tool. Elixir have several more differences?
It just have so much depth with concurrency. The actor model is amazing and easy to think about too.
I'm not entirely sure they overlap completely but the concurrency model is superb. I will probably only use Elixir for web application from now on (unless they don't have packages for certain API). Chris have made web development possible and many awesome people have contributed (POW package is just lovely).
Is it that hard to believe China haven't change?
Do you have a source for this? Is this audio compression for music production people or in general? I just can't see mp3 being out.