Ian Dworkin
Ian Dworkin
  • 37
  • 369 921
Very basic introduction to Bayesian estimation using R
datadryad.org/stash/dataset/doi:10.5061/dryad.8376 Using a simple general linear model as an example, this screencasts demonstrates some of the "canned" methods (I pre-built functions in R libraries) using Bayesian estimation and inference.
This is meant to provide a very basic overview of what results from MCMC can look like, and some simple diagnostics.
A screencast really going over the nuts and bolts to understand Bayesian estimation and inference,
as well as how MCMC works is still forthcoming!
Data can be found here
datadryad.org/stash/dataset/doi:10.5061/dryad.8376
zhlédnutí: 34 860

Video

Using R to fit regression models using maximum likelood
zhlédnutí 19KPřed 7 lety
Note (Sept 2019): New link to data datadryad.org/stash/dataset/doi:10.5061/dryad.8376 This screencast is a tutorial demonstrating how to fit simple general linear models (regressions and extensions) using maximum likelihood estimation. In it you will see how to write your objective functions, and how to use R's built in optimizers ( based on optim and wrappers such as mle() and mle2() in the bb...
Fitting simple models using Maximum likelihood using R
zhlédnutí 26KPřed 7 lety
Note (Sept 2019): New link to data datadryad.org/stash/dataset/doi:10.5061/dryad.8376 How to fit simple linear models (i.e. regression) using maximum likelihood by writing your own objective functions and using the bbmle() library (which provides wrappers for the optim() ). Surprisingly straightforward
Some old school plotting tricks in R (multiple plots on the same device)
zhlédnutí 319Před 7 lety
An old tutorial for some tricks for programming in R. In particular when you are trying to "over -plot" multiple things on the same device. This is with base plot tools.
Using the non-parametric bootstrap for regression models in R
zhlédnutí 13KPřed 7 lety
Note (Sept 2019): New link to data datadryad.org/stash/dataset/doi:10.5061/dryad.8376 This screencast continues the discussion and tutorial of using the non-parametric bootstrap for statistical inference, in this case for regression models (and the general linear model more generally).
Performing the Non-parametric Bootstrap for statistical inference using R
zhlédnutí 11KPřed 7 lety
Note (Sept 2019): New link to data datadryad.org/stash/dataset/doi:10.5061/dryad.8376 Using the R programming language to perform non-parametric bootstrap for statistical inferences, in particular generating confidence intervals. This includes the random variable ("pairs") bootstrap and the residual (fixed effect) bootstrap.
Permutation tests in R - the basics
zhlédnutí 10KPřed 7 lety
Note (Sept 2019): New link to data datadryad.org/stash/dataset/doi:10.5061/dryad.8376 Screencast demonstrating the basic approach to performing a permutation/randomization test using the R programming language. Demonstrates how to make statistical inferences using permutation tests.
Using the sample function in R for resampling of data - absolute basics
zhlédnutí 8KPřed 7 lety
Note (Sept 2019): New link to data datadryad.org/stash/dataset/doi:10.5061/dryad.8376 A short screencast introducing the sample() function in the R programming language which provides the basis for easy resampling of data, that in particular can be used (as shown in subsequent screencasts) for permutation tests and the non-parametric bootstrap.
simulations to understand relationships between variance, chi-square and F distributions
zhlédnutí 882Před 7 lety
Some students had some confusion between the relationships between variances (and estimates of variance) the chi-square distribution and the F distribution. Here is a short interlude screencasts that goes over these using some simulations in R.
Using monte carlo simulations to generate confidence intervals in R - part II
zhlédnutí 2,8KPřed 7 lety
Note (Sept 2019): New link to data datadryad.org/stash/dataset/doi:10.5061/dryad.8376
Using Monte Carlo simulations to generate confidence intervals in R
zhlédnutí 14KPřed 7 lety
Note (Sept 2019): New link to data datadryad.org/stash/dataset/doi:10.5061/dryad.8376 This demonstrates the basic approach to using monte carlo simulations to generate confidence intervals (sometimes called the parametric bootstrap) using the R programming language.
Using R to generate monte carlo simulations under null models
zhlédnutí 2,5KPřed 7 lety
Note (Sept 2019): New link to data datadryad.org/stash/dataset/doi:10.5061/dryad.8376 Generating simple monte carlo simulations in R under "null" models to aid in statistical inference. This is a basic approach to statistical inference (related to the parametric bootstrap discussed in the next screencast).
Using the R programming language for probability - part II
zhlédnutí 2,3KPřed 7 lety
The 2nd of two parts.
Using and exploring probability distributions using R
zhlédnutí 2,8KPřed 7 lety
This is a the first (of 2) screencasts on how to use and explore probability distributions using the R programming language. This screencast in the practical mechanics of using R, but does not include the more theoretical or conceptual background. This material is pretty fundamental if you are planning on using R for monte carlo simulations, maximum likelihood estimation or Bayesian estimation ...
General Linear Model - Simple model diagnostics
zhlédnutí 582Před 7 lety
This is an extension of the screencasts on the general linear model (statistics). Here we examine how well the model, data and fit behave with respect to several fundamental assumptions of the general linear model. Examples are in the R programming language.
General Linear Model - identifying & dealing with colinearity among predictor variables.
zhlédnutí 214Před 7 lety
General Linear Model - identifying & dealing with colinearity among predictor variables.
Multiple regression, multi-colinearity and interpreting partial predictors using R.
zhlédnutí 310Před 7 lety
Multiple regression, multi-colinearity and interpreting partial predictors using R.
Introduction to the general linear model using R - part III
zhlédnutí 524Před 7 lety
Introduction to the general linear model using R - part III
Introduction to the general linear model using R - part II
zhlédnutí 2,5KPřed 7 lety
Introduction to the general linear model using R - part II
Introduction to the R programming language - part II (old version)
zhlédnutí 210Před 7 lety
Introduction to the R programming language - part II (old version)
Introduction to monte carlo simulations using R
zhlédnutí 89KPřed 7 lety
Introduction to monte carlo simulations using R
Introduction to monte carlo simulations using R - The absolute basics
zhlédnutí 40KPřed 7 lety
Introduction to monte carlo simulations using R - The absolute basics
Review of the general linear model using R - part 1
zhlédnutí 3,1KPřed 7 lety
Review of the general linear model using R - part 1
Introduction to R part 10: Introduction to control flow
zhlédnutí 590Před 8 lety
Introduction to R part 10: Introduction to control flow
Introduction to Programming in R part 9: the family of apply functions
zhlédnutí 2,6KPřed 8 lety
Introduction to Programming in R part 9: the family of apply functions
Introduction to R part 8: Getting data into R
zhlédnutí 642Před 8 lety
Introduction to R part 8: Getting data into R
Introduction to R part VI: Writing Functions in R
zhlédnutí 2KPřed 8 lety
Introduction to R part VI: Writing Functions in R
Introduction to R part 7: Regular Sequences and Indexing in R
zhlédnutí 729Před 8 lety
Introduction to R part 7: Regular Sequences and Indexing in R
Introduction to R part V workspaces and how to use help in R
zhlédnutí 508Před 8 lety
Introduction to R part V workspaces and how to use help in R
Introduction to R part IV: Classes and Objects in R.
zhlédnutí 5KPřed 8 lety
Introduction to R part IV: Classes and Objects in R.

Komentáře

  • @vishalaaa1
    @vishalaaa1 Před 4 lety

    Please take a scenario and explain.

  • @ralphalbuquerque191
    @ralphalbuquerque191 Před 4 lety

    Hi Ian, Thank you very much for this video! I managed to apply the same code to my data after a lot of searching! However, when I include categorical IVs I get the following error: "argumento não-numérico para operador binário". I'm not sure why that error is in Portuguese while everything else is in English, but it translates as "non-numeric argument for a binary operator". Would you know anything about that? Thank you!

  • @guerschommugisho5569
    @guerschommugisho5569 Před 4 lety

    Is it possible to get your script?

    • @iandworkin1347
      @iandworkin1347 Před 4 lety

      I am hoping to put all of them in a github repo. These videos are all from a class I taught 8 years ago...

  • @guerschommugisho5569
    @guerschommugisho5569 Před 4 lety

    Dear Sir, I'm trying to run this program but I get this error message SimReg1 <- function(mod.input=lmx){ a=coef(mod.input)[1] b=coef(mod.input)[2] c=coef(mod.input)[3] d=coef(mod.input)[4] rse=summary(mod.input)$sigma y.sim <- rnorm(n=length(x),mean=a+b*x+c*x^2+d*x^3,sd=rse)$sigma lm.sim <- lm(y.sim~x) coef(lm.sim) } SimReg1() ``` Can you tell me how to solve this? I want to built confidence intervals for an adjuted polynomial curve

  • @surajbhagat2672
    @surajbhagat2672 Před 4 lety

    thank you so much for wonderful and easy to understand presentation, however, this is not for uncertainty purpose then would you like to enlighten how to do that, please?

    • @iandworkin1347
      @iandworkin1347 Před 4 lety

      Do you mean just for doing a monte carlo simulation for say a power analysis sort of approach (or simulations to develop theory)?

  • @hariadhikari6437
    @hariadhikari6437 Před 4 lety

    Is there a way to find those script

    • @iandworkin1347
      @iandworkin1347 Před 4 lety

      I hope to put them all up on github soon). I have not taught this class since 2014 (and these videos are several years older than that).

  • @valor36az
    @valor36az Před 4 lety

    Very helpful using different approaches and comparisons

  • @valor36az
    @valor36az Před 4 lety

    Thank you

  • @valor36az
    @valor36az Před 4 lety

    Great teacher

  • @belkisaltendji1249
    @belkisaltendji1249 Před 4 lety

    Hi, I need help please, I required for car pacadge, but the abswer was that there is no package , what should I d plz?

    • @iandworkin1347
      @iandworkin1347 Před 4 lety

      cran.r-project.org/web/packages/car/index.html

  • @vivianlondono2666
    @vivianlondono2666 Před 4 lety

    When i was doing the regression by the bootstrap method I had a big difference in some variables compared to the normal regression that may be due ?

  • @sfundomabaso3200
    @sfundomabaso3200 Před 4 lety

    Judging by your views, not so many people take States to this far

  • @shellyxiang6437
    @shellyxiang6437 Před 4 lety

    thanks so much for the videos and data, this help me a lot

  • @bobo0612
    @bobo0612 Před 4 lety

    Thanks a lot for your explanation.

  • @shwetalall6249
    @shwetalall6249 Před 4 lety

    Thanks! your video has been very useful but can you also guide me on how to introduce local minima and maxima for the input and output variable in the code.

  • @user-iq8ei9go3g
    @user-iq8ei9go3g Před 4 lety

    hello I wonder we can find the answer you posted for the exrecise.

  • @user-iq8ei9go3g
    @user-iq8ei9go3g Před 4 lety

    This is super helpful teaching!! Thank you very much!!

  • @emailtomurali2
    @emailtomurali2 Před 4 lety

    Loading required package: MCMCpack Warning message: In library(package, lib.loc = lib.loc, character.only = TRUE, logical.return = TRUE, : there is no package called ‘MCMCpack’

    • @iandworkin1347
      @iandworkin1347 Před 4 lety

      cran.r-project.org/web/packages/MCMCpack/index.html

  • @danielcarrillocolin724

    Hi, That is a nice video, i have a question. How do I interepete the results? Thanks

  • @kc_good_luck
    @kc_good_luck Před 4 lety

    Thank you! It's super clear!

  • @rps9167
    @rps9167 Před 4 lety

    It would be better if you put a link for part-1 under description

  • @aishamuhammad2543
    @aishamuhammad2543 Před 4 lety

    Please how can simulate a non homogeneous poisson process in r

  • @gustavofinoto2504
    @gustavofinoto2504 Před 5 lety

    Very helpful!

  • @daviddi7224
    @daviddi7224 Před 5 lety

    Thank you this was very helpful. How would you change the code for a small categorical variable (i.e., 0 = healthy, 1 = diagnosed disease) with only a small amount of confirmed diagnosed (i.e., 1s)? For example: There are 2000 participants and only 50 cases of cancer, 1500 healthy, and 450NAs.

  • @gbr4167
    @gbr4167 Před 5 lety

    thanks for sharing the skill, its very useful for my work^^

  • @boombangbang94
    @boombangbang94 Před 5 lety

    Hi Ian! Thank you for this tutorial. I have one question: when you generate y from rnorm with mean = a+b*covarite, is this mean a single number?

    • @iandworkin1347
      @iandworkin1347 Před 4 lety

      The way I have set it up in this, for each case yes. But the overall function is vectorized so it is generating the full vector of simulated responses.

  • @sethoduro1338
    @sethoduro1338 Před 5 lety

    Please I am trying to follow the bootstrap analysis. How can I get access to the data you used for the bootstrap estimate for confidence interval for the regression coefficients

  • @nguyentranvy2934
    @nguyentranvy2934 Před 5 lety

    Thanks Ian. I is so great to approach your video as I am begin to learn how to use bootstrap for my biodiversity analysis. I would like to follow your video and practice this then I hope I can apply this to my study. Could you help my to get your R script and data for this video? Thanks. My email is vychim@yahoo.com

  • @supreetisaha1630
    @supreetisaha1630 Před 5 lety

    Its very helpful. But where can I get the R-script of your programme?

  • @asunraychan
    @asunraychan Před 5 lety

    Thank you very much. I learned a lot from your video.

  • @tufleuddinbiswas7579
    @tufleuddinbiswas7579 Před 5 lety

    Happy to watch as it is started from the very beginning. Thank you so much.

  • @ninip4rk
    @ninip4rk Před 5 lety

    Fua sandra

  • @sann5146
    @sann5146 Před 5 lety

    Thanks for posting these videos! Just a quick comment regarding the variables at 1:00. I tried the example in Python and z didn't auto-update either there, when I changed x. When I re-assigned z = x + y ,then the new value of x was taken into account and z updated.

  • @hemmapermal532
    @hemmapermal532 Před 5 lety

    How to simulate seasonal data manually without arima package?

  • @mariakamran7442
    @mariakamran7442 Před 6 lety

    Could you kindly upload your code.

    • @iandworkin1347
      @iandworkin1347 Před 4 lety

      I am hoping to put them up on github soon. I have not taught this course for 8 years...

  • @stoptheangst
    @stoptheangst Před 6 lety

    i've been working with the data in matrix format. This program seems to rely on data frames ...any suggestions?

    • @NphiniT
      @NphiniT Před 5 lety

      Convert to DataFrames then!

    • @iandworkin1347
      @iandworkin1347 Před 4 lety

      Do you mean bbmle? I don't think it needs data frames. It could be the way I coded something.. I am hoping to put all scripts up and make links for them sometime soon. I have not taught this course in a long while.

  • @wnamal
    @wnamal Před 6 lety

    Thank you Ian for your time on showing this to us.

  • @yz1121
    @yz1121 Před 6 lety

    Hello Ian, may I have the command or any other supporting documents to run Bayesian spatiotemporal analysis for count data, please?

  • @Mario94177
    @Mario94177 Před 6 lety

    Hello Ian, do you know any packages in RStudio that would enable me to perform a laplace mle fit on some data and then plot the fitted line along with the empirical data?

  • @Emmyb
    @Emmyb Před 6 lety

    Brilliant thank you!

  • @Emmyb
    @Emmyb Před 6 lety

    Thank you sooooo much!!!!!!

  • @ShooprDoopr
    @ShooprDoopr Před 6 lety

    Thank you for this tutorial series! This has helped a lot!

  • @rghouse530
    @rghouse530 Před 6 lety

    what an awesome program, and easy to follow beginner's tutorial! I was fortunate enough to have stumbled upon this right after I learned how to use TPS programs. I had all of my landmark data ready to go and was able to follow these three videos to analyze the shape differences between male and female Chinook Salmon, as well as other fishes in our system; improving data analysis/fish identification techniques. Hopefully phylogeny of Trout next (?).... more tutorials please!!!!

  • @soumikchatterjee3996
    @soumikchatterjee3996 Před 6 lety

    Really helpful. Thanks

  • @hx3tube
    @hx3tube Před 6 lety

    boring talk

  • @yiyingw6506
    @yiyingw6506 Před 6 lety

    Thank you! These videos are very helpful. If you also watch the videos in order like me, be aware that the previous videos of this one are "Using Monte Carlo simulations to generate confidence intervals in R" and its part 2.

  • @sebastiankian693
    @sebastiankian693 Před 6 lety

    This video literally just saved my ass with my assignment, thank you!

  • @sciencescience9469
    @sciencescience9469 Před 6 lety

    MorphoJ can set up for window

  • @tomweinandy
    @tomweinandy Před 6 lety

    Thanks also! Any chance you could share the script file?

    • @iandworkin1347
      @iandworkin1347 Před 6 lety

      Sorry about this, the site that had all of the scripts up is gone. I am slowly migrating everything to github. Some of the older ones (that pertain to the basic R stuff) should be here. github.com/DworkinLab/CSE845_R_tutorials If you want the more advanced stats one, email me and I will send you the scripts. I don't teach these courses anymore, so I am not checking this very often.

    • @rodrigopierott
      @rodrigopierott Před 4 lety

      @@iandworkin1347 Hi, I couldn't find the script, could you point me where exactly is in your repository? Thank you very much!

  • @subhadeep1024
    @subhadeep1024 Před 7 lety

    what does it mean when a vector is used as mean instead of single value in rnorm?

    • @iandworkin1347
      @iandworkin1347 Před 6 lety

      If you do something like rnorm(10, mean = c(5, 50)) you will see that it will alternate in generating random numbers either with a mean of 5 (odd outputs) or 50 (even outputs).