you're my role model! I've learned so much about the tidymodel package from you! Thank you my dear mentor.
@gecarter53
4 жыл бұрын
This was a great and timely video. It gave us an opportunity to improve our data analysis skills, and at the same time educate ourselves about some of the history of people of color. As a result of the Black Lives Matter movement, many social media content providers have started incorporating such histories in their content, and should continue to do so regarding people of all colors. Thanks again for a great video and hoping to see more from you.
@joel09878
4 жыл бұрын
Thanks! I will definitely try tidymodels for imputation in future, it's so clear and logical. The dataset was also fascinating and valuable. Much appreciated.
@rrrprogram4704
4 жыл бұрын
Awesome Julia... can you please make videos on each package in Tidymodels... how each package play its role ..... There are 1000 of options ... But, going by 80-20 rule... 20% of the general application will be used 80% of the time... with this in mind... can you please make series??? Bcoz I guess there is NO channel on youtube which elaborately talk on tidymodels
@JuliaSilge
4 жыл бұрын
You can get a high-level view of what the different packages do here: www.tidymodels.org/packages/
@maksim0933
4 жыл бұрын
thanks for the lesson! please more videos about missing data imputation) very appreciated)
@danielalvarezmd
4 жыл бұрын
Hi Julia, first congratulations on your videos. They are an excellent learning material!. I love the color scheme you use in Rstudio. Is it a specific theme or did you customize it?
@JuliaSilge
4 жыл бұрын
It is one of the themes available from the rsthemes package, I believe Oceanic Plus: github.com/gadenbuie/rsthemes
@lucianobatista4554
4 жыл бұрын
I also really like this color scheme and didn't know the rsthemes package, thanks Julia.
@b4r3sGT
4 жыл бұрын
@@JuliaSilge Your are the best. Thank U!
@mmbodnar
3 жыл бұрын
Thanks for the tutorial. Do you use a custom theme for R-studio?
@JuliaSilge
3 жыл бұрын
I believe I use one of the themes from rsthemes: github.com/gadenbuie/rsthemes
@pritomroy2465
3 жыл бұрын
Thank you for the great video. I finished all your videos. Please give me some more resource in r tidymosels, tidyverse like this in youtube,,,
@deependradhakal7635
4 жыл бұрын
Hi! Very smooth workflow there you have. One point though, since you construct a linear model finally and fit and interpret, I notice that R squared values are too low. Doesn't this mean the model is not very useful ?
@JuliaSilge
4 жыл бұрын
Definitely not very useful if you want to predict the year from gender and age, for sure! R squared is a measure of how much variance in the year is explained by gender and age, which as we would probably expect is very low here.
@barnettc
Жыл бұрын
I know this is an older video, but I'm hoping someone can help me understand the gg_miss_unset() plot at 23:53. How is there more rows missing age, height, and gender (114) than missing just age (109)? Thanks
@JuliaSilge
Жыл бұрын
It is unusual to see something like that (typically there are more rows missing single values than multiple values) but it certainly doesn't have to be like that. In this case, more people did not have their age, height, and gender recorded that did not have only their age recorded. Remember that these are disjoint sets, not overlapping sets. You can read more here: naniar.njtierney.com/articles/naniar-visualisation.html#exploring-patterns-with-upsetr
@barnettc
Жыл бұрын
@@JuliaSilge Thank you! I was assuming they were overlapping sets. It makes sense now. Thank you for putting out this content. I love your videos (and books)!
@Hollix00
4 жыл бұрын
Julia, thanks for the great video. One question. In my field it is usual, to impute several data sets to reflect some kind of variance and then to combine the reults of the model by the so-called Rubin-formula. Can you tell me if this is possible in the approach that you presented? Many thanks
@JuliaSilge
4 жыл бұрын
You can check out the types of imputation currently implemented in tidymodels here: recipes.tidymodels.org/reference/index.html#section-step-functions-imputation You certainly can create multiple imputed datasets, then fit a model to each imputed data set, then combine or pool the estimates from each model. You can do that using dplyr and the with and between imputation variance, etc (Rubin's rules) but you might be interested in checking out tidyposterior: tidyposterior.tidymodels.org/
@Hollix00
4 жыл бұрын
@@JuliaSilge Julia, if I may ask a follow up question. I applied your code presented in the video but repeating the imputation results in exactly the same imputations (hence: no variance). Do get something wrong?
@charithwijewardena9493
4 жыл бұрын
Hi, thank you so much for these videos. It's by watching your videos and reading R4DS over the last few months that I've learnt R and machine learning! I have a question about handling missing values and would appreciate any insight you could share. For my work I'm trying to predict a patient's test_result_B based on test_result_A and other predictors like age, gender etc. However I'd also like to take into account previous results for test A. But obviously not all patients would have done test A previously. It doesn't seem correct to impute a patients previous results based on a mean or even knn. My thought was to perhaps encode previous_test_A as say -100 if absent else the actual result. Is this the right approach? And if so does this approach work for any model? Much gratitude for any advice or pointers to helpful resources.
@JuliaSilge
4 жыл бұрын
Hmmmm, I think a little more information could be helpful to give you a good answer here. Could you put together a little more detail and post on RStudio Community? It's a little easier to answer more detailed questions like this there: rstd.io/tidymodels-community
@charithwijewardena9493
4 жыл бұрын
@@JuliaSilge sure, thank you. :-)
@moujib_jarray
3 жыл бұрын
Nice..are u unsing Python ?
@flamboyantperson5936
4 жыл бұрын
Hi, Please suggest me some best machine learning python channel?
@JuliaSilge
4 жыл бұрын
I really like Rachael Tatman's live coding in Python on Twitch: www.twitch.tv/rctatman
Пікірлер: 28