- Видео 71
- Просмотров 92 592
BIOS 6611
Добавлен 23 июл 2021
The RUclips channel for course videos for the BIOS 6611 Biostatistical Methods I course at the University of Colorado-Anschutz Medical Campus taught by Dr. Alex Kaizer (Fall 2020-Fall 2022) and the BIOS 6618 Advanced Biostatistical Methods I course (Fall 2023 onward). This course is an introduction to applied biostatistics, including probability models, simulation, sampling distributions, hypothesis testing, resampling methods, linear regression, and the use of the R (and some SAS). Additional content, including slides for the videos, can be found at www.alexkaizer.com/bios_6611 or www.alexkaizer.com/bios_6618.
Bayesian Linear Regression
While much of our semester has focused on frequentist methods, we can also implement regression models in the Bayesian framework. In this lecture we revisit our multiple linear regression model lecture's data set and MLR to refit it from a Bayesian perspective. We explore 3 different prior specifications ("noninformative", "informative", and poorly specified) to see how the posterior does (or does not) differ. All examples are coded in R using the brms package.
A video for the Advanced Biostatistical Methods I (BIOS 6618) course in the Department of Biostatistics and Informatics at the University of Colorado-Anschutz Medical Campus taught by Dr. Alex Kaizer. Slides and additional material...
A video for the Advanced Biostatistical Methods I (BIOS 6618) course in the Department of Biostatistics and Informatics at the University of Colorado-Anschutz Medical Campus taught by Dr. Alex Kaizer. Slides and additional material...
Просмотров: 674
Видео
Intro to Bayesian Methods
Просмотров 1,1 тыс.Год назад
A brief introduction to the concepts and terminology behind Bayesian methods in preparation of our example lecture for using Bayesian regression for linear regression. We introduce the concepts of Bayes' theorem, the idea of MCMC and diagnostic plots/numerical summaries for the estimation, different quantities to estimate from the posterior distribution, and a comparison of using the brms packa...
Advanced Bootstrap Topics
Просмотров 191Год назад
This lecture builds on our exploration of nonparametric case-resampling bootstraps by exploring more advanced confidence interval calculations (BC, BCa, and t-studentized) and different bootstrap strategies (parametric, smoothed, residual, wild). A video for the Advanced Biostatistical Methods I (BIOS 6618) course in the Department of Biostatistics and Informatics at the University of Colorado-...
Bootstrap p-values
Просмотров 683Год назад
This lecture introduces how we can use bootstrap resampling to calculate p-values by resampling from the null distribution of our statistic. We cover different p-value approaches and trade-offs for calculating one- and two-tailed p-values with our bootstraps. A data analysis example is provided at the end with a comparison of multiple statistics, null strategies, and p-value approaches. A video...
Splines in Linear Regression
Просмотров 1,5 тыс.Год назад
A brief introduction to the motivation and approach behind modeling non-linear trends in linear regression with splines in comparison to polynomial regression. B-splines and natural cubic splines are introduced with some brief technical details before exploring examples of fitting and changing the shapes of splines in an NHANES data example. A video for the Advanced Biostatistical Methods I (BI...
Segmented Regression
Просмотров 1,8 тыс.Год назад
A brief introduction to the motivation and math behind segmented regression and how to implement it in R. We explore a data example using NHANES data to fit a breakpoint to SHBG across age. The code example illustrates how to evaluate for multiple breakpoints and to evaluate the model fit. A video for the Advanced Biostatistical Methods I (BIOS 6618) course in the Department of Biostatistics an...
Quantile Regression
Просмотров 396Год назад
A brief introduction to the motivation and approach behind quantile regression and an example of its application in R. The example includes calculation of p-values and bootstrap percentile CIs. A video for the Advanced Biostatistical Methods I (BIOS 6618) course in the Department of Biostatistics and Informatics at the University of Colorado-Anschutz Medical Campus taught by Dr. Alex Kaizer. Sl...
Power and Sample Size
Просмотров 4933 года назад
Now that we've covered the null hypothesis significance testing (NHST) framework and how we use it for statistical inference, we can move to one of the most feared and misunderstood topics in BIOS 6611...power and sample size calculations! In this lecture we will derive closed form solutions for power calculations for a two-sided, one-sample Z-test. There are 5 important quantities in almost an...
A Brief Introduction to Generalized Linear Models
Просмотров 2,8 тыс.3 года назад
As one of our final videos for BIOS 6611, we introduce the concept of the very flexible generalized linear model. The linear regression framework we've been working in this semester is a subset of a much larger world of regression models that can be considered (e.g., logistic regression, Poisson regression, etc.). The purpose of this lecture is to lay some of the groundwork for what you'll be g...
Linear Regression: MLE for Methods of Estimation
Просмотров 7 тыс.3 года назад
Most of our semester in BIOS 6611 has focused on the ordinary least squares approach to estimation and derivation of our regression coefficients. However, we can also use maximum likelihood estimation to derive our beta coefficients and other quantities of interest (and, spoiler, the beta coefficients are identical to OLS!). The MLE approach is more general, however, and will be useful as you t...
Linear Regression with Matrices in SAS
Просмотров 4923 года назад
In addition to using procedures like PROC REG, we can also use PROC IML to directly code matrices and implement our own regression models. In this lecture we work through using PROC IML and compare results to our "by hand" calculationss from a previous lecture and the output from PROC REG. A video for the Biostatistical Methods I (BIOS 6611) course in the Department of Biostatistics and Informa...
Linear Regression with Matrices in R
Просмотров 2 тыс.3 года назад
Beyond using the existing functions in R for linear regression (e.g., lm, glm), we can also code our own matrices to implement a linear regression model. In this lecture we walk through an HTML document with R code and compare to existing functions. A video for the Biostatistical Methods I (BIOS 6611) course in the Department of Biostatistics and Informatics at the University of Colorado-Anschu...
Linear Regression with Matrices
Просмотров 2,5 тыс.3 года назад
Most of our coverage of linear regression has focused on an algebraic approach to deriving the coefficients and estimating parameters. However, we can accomplish the same feat by using matrices to represent all our coefficients, variance estimates, and other regression-related quantities. In this lecture we describe how linear regression can be approached using matrices. An example of the calcu...
Multiple Linear Regression: Outlier, Leverage, and Influential Points
Просмотров 3,4 тыс.3 года назад
An observation in our regression model may seem to depart from the rest of our sample, which may or may not reflect a true phenomenon. In this lecture we discuss how to identify outliers, leverage points, and influential points in our regression analysis and how to address them. A video for the Biostatistical Methods I (BIOS 6611) course in the Department of Biostatistics and Informatics at the...
Multiple Linear Regression: Variable Selection
Просмотров 1,7 тыс.3 года назад
When there are multiple variables to select from that may be of potential importance in a regression model, we may have to select some smaller subset due to a smaller sample size (e.g., smaller n than p) or to try and achieve a more parsimonious model (i.e., fewer variables may be easier to interpret and disseminate). In this lecture we highlight some variable selection strategies, as well as d...
Model Selection in Linear Regression
Просмотров 1,2 тыс.3 года назад
Model Selection in Linear Regression
Multiple Linear Regression: Interactions (Effect Modification)
Просмотров 1,3 тыс.3 года назад
Multiple Linear Regression: Interactions (Effect Modification)
Confounding and Precision Variables in Linear Regression
Просмотров 8693 года назад
Confounding and Precision Variables in Linear Regression
Kruskal-Wallis Test (Nonparametric ANOVA)
Просмотров 1733 года назад
Kruskal-Wallis Test (Nonparametric ANOVA)
Multiple Comparisons and the False Discovery Rate
Просмотров 4,8 тыс.3 года назад
Multiple Comparisons and the False Discovery Rate
Multiple Linear Regression: Inference on Independent Variables
Просмотров 5363 года назад
Multiple Linear Regression: Inference on Independent Variables
Multiple Linear Regression: Diagnostic Plots and Multicollinearity
Просмотров 5103 года назад
Multiple Linear Regression: Diagnostic Plots and Multicollinearity
Multiple Linear Regression Introduction
Просмотров 2133 года назад
Multiple Linear Regression Introduction
Transformations to Remove Heteroscedasticity and Address Non-Linearity
Просмотров 1,5 тыс.3 года назад
Transformations to Remove Heteroscedasticity and Address Non-Linearity
Thank you so much! Concise, straight to the point and easy to understand 🎆
Thank you very much for this nice explanation!
16:52 Bootstrap sampling IS based
that intro is something else lol. great video thanks so much better at explaining than my professor.
This is very good. i hate smart people, though. LOL. Humbling to see how fluid they re with things I struggle with.
Thanks! Very helpful and understandable!
😊😊thanks for excellent explanation
excellent thank you
thank you! I needed this.
I have a doubt. The segmented package returns only one coefficient of determination (Multiple R-squared). Does this Multiple R-squared consider the R-squared of the two models if I have one breakpoint or tree models if I have two breakpoints? Is there a way to know the R-squared of each model fitted? Thank you
While I can't speak to all possibilities from the package or the ability to derive separate R-squared values for each segment, the output from the package should represent the overall variability in our outcome explained by the given model. So, if we have 2 breakpoints, the R-squared included in the output is the overall summary for that model having 3 segments. One option for calculating an R-squared for each segment could be to estimate the cutpoints, then fit individual linear regression models restricted to the data within each segment. However, I am not sure that this information is as useful as the overall R-squared from the segmented package, since it is not clear if one segment had lower R-squared if that is important relative to the overall performance of the given model.
I love you, thx.
Thank you this was very usefull! i know that you adviced for the usage of confidence interval to test for significant slopes. However, if i decide against that and use an anova on the segmented model it becomes somewhat tricky to interpret it... And i cant find resources regarding anovas on segmented models. Do you have any idea where i can get some insights?
BAM! Best intro ever!
very well explained.
where can i find more content on generalised linear model
This is helped me so much in my Ege University Final Exams, Thank you very much!
Thank you...Awesome and Clear explanation
This is really excellent, really useful
Simply awesome and lucid explanation
how one would go a bout simulating a stopping time itself as a standalone random variable ? and are there any applications in which we would be interested in simulating such r.v. and under which "statistical"conditions such exercise would be a useless computations ?
hello sir. Can I have the data used in the example? I trying to do it with R
The StatQuest is strong with this one
Hello, in the derivation for the intercept, for the Y_bar, I think it's also a constant. So Var(Y_var) will be zero too.
the intro song caught me off guard ngl
Tnx! Helped me a lot. :)
BIOS 6611 is filmed in front of a live audience
Excellent video!
guess they should have named it Fisher's TEA Test nah we glorified that eugenicist enough :)
Great video, thank you!
Those t values are palindromes if you ignore the negative sign and the decimal point
Absolutely the best in depth explanation of the Permutation test w H0/H1 w P-Value, from anywhere on the web (and I had seen some 20-30 versions. Also not much on this type of thest around. Unfortenately I could not reproduce it in R studio (errors like: 'Error in sample.int(x, size, replace, prob) : cannot take a sample larger than the population when 'replace = FALSE'' or 'invalid number of 'breaks'', or 'object 'observed_diff' not found'... spoiled the fun for me. But even without it just conceptually it helpd great deal. Before it was like a black box- now is more of a logical box. Thank you! ( i come from backg of no math and this was first venture into R)
Thank you so much. I was actually looking for that you solved. Great work
Execellent video with great example.......
Thanks, this video was very helpful.
Thank you! Perfectly explained step by step for calculating sample size
Perfect. Thank you very much!
Very nicely explained! liked the video!
5:39 - various piles of gold tumbling