/stat/ - statistic general

[Anonymous No.9649936]

discuss statistic shit. machine learning, biostatistics, data science and so on

[Anonymous No.9649987]

Entropy/information theory is the best approach to prob/stats.

[Anonymous No.9649990]

Who else stats and computer science double major master race?

[Anonymous No.9650013]

Why didn't I major in Stats instead of math. So much more cmfy

[Anonymous No.9650014]

>>9649990
took all my gen eds in 2 years at a community college.
now a stats/app math double major with cs minor. that, freshman phys and french is all i have to take now

[Anonymous No.9650052]

>>9650014
>app math

Do pure math, you need it rather than the watered down stuff to do anything meaningful.

>cs minor

If you want to take cs courses, take graduate courses.

[Anonymous No.9650066]

>>9650052
The difference between app and pure undergrad here is algebra II and topology. App math still have to take analysis I&II and algebra I

[Anonymous No.9650078]

>>9650052
>>Pure unemployment major

[Anonymous No.9650377]

>>9649936

Are there any good online references for data science? After I'm done this semester I was going to work through this:

http://www.datasciencecourse.org/

But I'm not entirely certain one semester's worth of data science is all I need to be a functioning data scientist.

[Anonymous No.9650564]

>>9650377
>But I'm not entirely certain one semester's worth of data science is all I need to be a functioning data scientist.
That depends how you define "data scientist."

[Anonymous No.9651613]

>>9649936
Anyone here into econometrics? I would like to hear the kinds of questions it may answer. Also if someone has ongoing research in this I'd like to hear about it.

[Anonymous No.9651683]

>>9650066
What textbooks do you use for algebra and analysis?

[Anonymous No.9651818]

>>9651683
for analysis, its rudin

the algebra book is dummit and foote

[Anonymous No.9652159]

>>9651613
I use econometrics for class and my internship all the time, ama

[Anonymous No.9652167]

Exponential-kun is best girl

[Anonymous No.9652316]

bump for a great, cozy and underappreciated field

[Anonymous No.9652330]

>>9650078
undergrad stats is just as unemployable.

[Anonymous No.9652351]

How useful is bayesian statistics

[Anonymous No.9652645]

Is it still possible to get into grad school for stats or have pajeets taken this too

[Anonymous No.9652943]

>>9652351
Very underrated, Bayesian network analysis is the next big thing

[Anonymous No.9653004]

>>9652159
Hmmm I don't know much, it's something that does kinda intrigue me though. I mainly just wanted an abstract of how you are applying it in your internship for instance. What software do you use? And does it involve machine learning for regression analysis?

[Anonymous No.9653124]

>>9653004
Well it's mainly suited to estimating relationships with noisy data, which is why it's usually used for quasi experimental data. I use STATA mostly but R and Python are also pretty commonly used. Most academic studies don't involve methods with machine learning but econometric methods are used with machine learning in finance all the time

[Anonymous No.9653803]

>>9652645
Yes, many substantive programs have a stats specialization that, depending on how far you're willing to push it, can be just a rigorous as a statistics program.

[Anonymous No.9653867]

>>>/pol
Statistics is racist.

[Anonymous No.9653899]

>>9652330

[Anonymous No.9653943]

>>9653867
how?

[I Go !2HARVARdHk No.9654271]

>>9649990
>computer science

[Anonymous No.9655620]

Bump because stats is underrated

[Anonymous No.9655863]

Unsure if I want a phd in stats or probability. My goal is to work for the govt, maybe at the cdc or in public health. But I havent decided specifically yet. So far Im in my undergrad, stats+app math major with a bio minor.Very comfy atm

[Anonymous No.9656149]

can anyone actually define what a "significance level" is or what a "confidence interval" is?
I feel like everybody is able to plug and chug but no one really has any idea what's going on

[Anonymous No.9656214]

Is there a Pearson-like coefficient, but where 0 is no correlation and 1 is max correlation, positive and/or negative?

[Anonymous No.9656513]

>>9656214
The absolute value of your coefficient [-1,1] would map to [0,1] and have the desired properties

[Anonymous No.9656524]

I have got some background in analysis, discrete and linear algebra (just default undergraduate courses I guess), we have a statistics/probability theory course where I slacked off. We are now at uniformly most powerful tests, and considering I barely remember formula for binomial distribution, I am properly fucked.
What are the must-have probability theory/statistics books?

[Anonymous No.9656532]

Who /actuary/ here?
Learn stats, feel like a Wolf of Wall Street type, and command dominion over life and death.
Shit’s cash.

[Anonymous No.9656651]

Is clustering a meme?

[Anonymous No.9656664]

Made a thread for this without seeing this general haha--

What is the probability that the skill level for an open demographic of any sport (inclusive of amateur to professional) be evenly distributed across every country?

i.e. 20% suck / 50% suck less / 20% can dribble and shoot moderately / 7.5% are good / 2% are really good / .5% are professional

What are the odds that across every region, the difference in percentiles were 0-5%?

Likely or a statistical impossibility?

My rationing currently is that there's always going to be varying factors across different lands so the skill distribution should never be statistically similar.

Thoughts?

[Anonymous No.9656671]

>>9656524
fucked for what?
>>9656532
>actuary
>wolf on wall street
those guys are economists faggot

[Anonymous No.9656697]

What's a good distribution for modeling lifespan for a species where infant mortality is pretty high but then the longer an individual is alive the higher the probability that they will continue to live (up to some threshold)?

[Anonymous No.9656777]

>>9656697
some variant of the red one?

[Anonymous No.9656802]

Just got to know Survival Analysis. This shit is so cool.
I can't believe I am studying something with names like " Hazard function ", " life table " and " probability of death "

[Anonymous No.9656847]

>>9656671
I wanted to learn data analysis/statistics to use it later in Python/R, and would like to be somewhat proficient in it. I fail to comprehend most concepts we passed.

[Anonymous No.9656868]

>>9652943
maybe for medicine lmao

[Anonymous No.9657135]

>>9656697
Why wouldn't you just use the Cox proportional hazards model?

[Anonymous No.9657206]

>>9656532
Studying this, have you qualified?

[Anonymous No.9657221]

>>9656532
you wouldn't happen to be a 2nd semester freshman with mediocre math ACT/SAT scores, never really the top student at anything in high school but you could at least compete with the lower end of the best students, would you?

[Anonymous No.9657378]

Is "Probability and Statistics" by DeGroot and Schervish (4th ed.) any good? It's the book for my intermediate probability course in the Fall. We're covering the first 7 chapters but the course doesn't have a detailed schedule of topics and I'm too lazy to download a 200MB book just to check its table of contents.

[Anonymous No.9657481]

Why are statistics departments at every school either non-existent or extremely small? You can even look at the well-known statistics departments (e.g. CMU, Stanford, UCLA, ...) and their departments are still very small. How do statistics majors survive? How can we save statistics? Why isn't it more popular with all the ML hype at the moment?

[Anonymous No.9657535]

2nd yr stats major here. Shit is so cash, the more you get into it the harder it is to escape the abyss.

[Anonymous No.9657544]

>>9657221
2nd semester junior, great sat and act scores, two summers of actuarial internship experience. Get fucked.

[Anonymous No.9657558]

>>9657544
Where do you go to school?

[Anonymous No.9657706]

>>9657481
Do you really want another field to saturate, anon? We're comfy and in-demand as it is.

[Anonymous No.9657715]

>>9657706
Yes. There should be exploration for undergraduate statistics at every university. It's massively underrated and too many people want to get into ML without the statistics background, which is vital.

[Anonymous No.9657730]

>>9657715
Nobody's keeping newfriends from applying to any stats dept, most people are just art major brainlets.
As for the fags that want to get into ML without any a priori knowledge of statistics, they can get fucked.

[Anonymous No.9657941]

>>9657481
For a few reasons

1) Statistics is a very dry job. Little creativity unless you go into research or have a nice boss. 90% of stats grads will be number crunching at a desk. Thats not appealing to most people.

2) Statistics isnt difficult, most scientists pick it up on their own. So there isnt a ton of demand for statisticians in science fields. If science/math interested students arent majoring in it, then who is? Theres little reason to take stats for the type of student who would be interested in it.

3) Almost no advertising. And why would there be? Its not glamorous.

Btw if youre a stats major, you should double major in a field you intend to enter. Math, bio, engineering, computer science. something

[Anonymous No.9657960]

>>9657941
>Statistics isnt difficult, most scientists pick it up on their own. So there isnt a ton of demand for statisticians in science fields.
Sure, but a lot of those 'scientists' use basic statistics and very basic data visualization. Not many people appreciate how deep statistics can go.

I'm a computer science major by the way, planning on minoring in statistics. It's completely relevant simply because of the probability, time series, and rigorous mathematical statistics. I've already taken AI and data science courses for my major, so I might as well take the rigorous/uncommon stats courses that might be applicable in the future.

What do statisticians even do if it's not fun? I thought a lot of you guys became quants or something making 200k a year.

Am I fucked for getting a stats minor? Will it be useless in the future?

[Anonymous No.9657990]

>>9657941
>>9657960
>Am I fucked for getting a stats minor? Will it be useless in the future?
I never understood this. How is it a useless major when a lot of websites say it has a good job market for the future?

[Anonymous No.9658084]

>>9657960
>Am I fucked for getting a stats minor? Will it be useless in the future?
What do you mean? Its not bad to get it if you want. It doesnt worsen you in any way.

>What do statisticians even do if it's not fun? I thought a lot of you guys became quants or something making 200k a year.
90% of stats undergrads work basic data analysis jobs. start around 60k and you might make 100k in 10 years

quants major in math or stats then go on to get probability phds for the most part. they dont really use statistics except to organize large sets of data

>>9657990
>How is it a useless major when a lot of websites say it has a good job market for the future?
what websites? i wouldnt trust any site. professors are far more knowledgeable on the nuances involved in the stats market.

for example, the boom in big data is going to quickly collapse as regulation gets written to curb it. big data played a MAJOR role in the recent election of the US president. expect major changes.

a less intense thing is going to happen in biostats/health. pharma companies have free reign right now. but the public is getting fed up with them and demanding regulation. i imagine the demand for biostatisticians will decrease as drug companies stop being allowed to sell the same drug under new names. plus there is calls for the govt to compete with private companies which will decrease available jobs

[Anonymous No.9658136]

Currently a math major in the first statistics class for scientists and engineers at my school. Anyone mind telling me what the difference between mathematical biology and biostatistics is?

[Anonymous No.9658241]

>>9652167
Having OpenBugs flashbacks

[Anonymous No.9658246]

>>9653867
Sam Harris v Ezra Klein

[Anonymous No.9658257]

>>9658136
Biostatistics is using statistical models, analyzing data and designing experiments. Biomath is using math models for biological phenomena

[Anonymous No.9658630]

How do I learn R real well for bioinformatics & biostats purposes

[Anonymous No.9660598]

>>9656149
What? A confidence interval is just a range of values away from the average that are most likely to occur (typically with a 95% probability - or two standard deviations). Anything outside of the confidence interval is considered unlikely to occur. What isn't to get here?