Linear Regression Summary in R

Linear Regression Summary in R
Linear regression is an essential tool in R, but the output can be a little difficult to interpret. In this video, I walk you through the basics of the output and let you know what type of values to look for when you're fitting a linear model to your data.

[@nikolescobedo7024]

learn more in 10min than in one semester of stat THANK U SOO MUCH

[@Annielytics]

This is the best explanation I've seen to date. Thanks for focusing more on what the numbers mean and the conclusions we can make about our models and less on the underpinning formulas.

[@analyst42]

WOW! This is easily the best explanation of the summary statistics for a linear regression model that I've ever encountered. Thank you so much!

[@kennedygolfhead4356]

MY GOD!! IT WAS SOOOOO HELPFUL!! Best explanation ever in just 10 minutes!!!!! How life saving this is for all academics and data scientists!! Would you please please consider going over the summary of Linear Mixed Modelling??? THANK YOU A MILLION!!

[@georgie66549]

I wish there was a love button for this video. Thank you so much!!! I read through my lecture and still had no real clue what I was looking at in R. My understanding is so much clearer now

[@thesoaringdividend3830]

LOVED THIS video! Way better than how my Data Mining professor explained it 😅

[@rohitekka2674]

Thank you for the wonderful explanation. This video really felt like light at the end of the tunnel. An absolute enlightenment.

[@jasperembiricos2704]

You just saved my undergrad degree and I'm not kidding

[@reubenbrown1995]

Brilliant, compressed a week of lectures into 10 minutes!

[@rasmusnordman6764]

Absolutely incredible video, thank you so much!

[@luismi8936]

Thank you man, so accurate. So much information well explained. You're amazing

[@CamViesky]

Absolutely brilliant, clear, and concise explanation. Thank you.

[@ianleboo]

Dataslice is just wow... Precise, informative and accurate. Good work👏 Much love from 🇰🇪

[@curasvitas96]

oh my gosh this is one of the best explanations I have seen. thank you very much!! I love the part you touched on each of the indicators of the summary() code mean. thank you!!

[@edadila]

this was all i was looking for. perfect explanation, thank you!

[@NamNguyen-kp1xu]

Such an underated channel, clean explanation and straight to the point !

[@annazheng3391]

So clear and helpful!! Answered all my questions in 10 mins. Thank you!

[@marcodomenicoolivi352]

in 10 minutes, you have explained the linear regression much better than my professor

[@krisztinachristmon9978]

Thank you so much! Finally was able to make sense of the R output and interpret my data.

[@dodgecarlincila879]

Thanks for the visual examples, very helpful. Can't wait for your Machine Learning videos!

[@emmy8133]

This really solved my question, thanks a lot.

[@HuongGiangNguyen-qt3sm]

Thank you! Very well explained.

[@CanDoSo_org]

Thank you. Nice and clean.

[@amaebarnes]

I was just trying to brush up on my econ/statistics degree because it's been 5 years since I was in university. I just realized how much I've forgotten 😳 😬

[@Kenkoopa44]

Excellent!!!! Thanks so much!

[@ihsan3700]

THANK YOU. Such an amazing lecture .

[@shuvhamdigitalacademy3228]

Kindly make some videos on multiple regression analysis and interpretation.

[@johners2006]

Very clear and helpful. Pace was good. Thanks for doing it.

[@nextlevel6825]

Incredible 🚀🚀

[@malinkata1984]

Thank you. Great explanation. :)

[@kelliegadeken8481]

Wow, so clear and easy to understand. Thanks!

[@brazilfootball]

Great video, thank you! Any chance you could make one for the summary output of other types of regression models (Logistic, neg. binomial, Poisson, etc.)?

[@danielsummers9345]

Very helpful, quick and easy to understand!

[@dataslice]

Thanks! :)

[@zakariya_am]

Thank you

[@genevieveemefaasare8352]

Thanks so much.

[@tagamag]

Very well done. Please do a few such videos on Stepwise regression, Logistic regression etc.

[@jives.]

thank you dataslice gang

[@ahmed007Jaber]

this is superb and clear. thank u for this

Excellent! Thanks for sharing.

[@chinthakajayarathne]

Thank you for video. Much helpful !!

[@asseflas]

This was great thanks :)

[@darkrillen]

well said

[@cezreycor]

pretty cool stuff man, many thanks for the clear explanation!

[@amirmirou8037]

Dude thanks a lot!

[@harentongaray3443]

So, if the P value is less than. 0.5 the model is significant and therefore we void the null hypothesis?

[@marvinschumann6832]

Great video my man!!! Thank you so much!

[@aishwariyagupta619]

How do we know residuals are normally distributed?

[@abdullahalayed5276]

Very good. Appreciate the effort.

[@Quetzal00358]

This was great! Have you made a video where you also include categorical variables?

[@hashemfathi1646]

spot on.... just one question, what does the " on 2" mean in the F statistic part?

[@dataslice]

Good question -- that's how many x variables we used for the regression

[@jac6003]

Great channel! Thanks!!

[@nooberinho]

Very helpful, thanks

[@CaseySchacher]

How do I access the second video talking about diagnosing regression models? You mentioned that were going to make the new modeling/diagnostics video towards end of this one. Really appreciated this video!

[@dataslice]

Yes, I'm hoping to release a video covering the linear regression plots soon and then potentially more regression videos down the line!

[@kennylee6768]

So what is the regression equation?

[@aishwariyagupta619]

Hi, can we determine the sample size from this output?

[@apurbabiswas8072]

Waiting for the 2nd video of this topic

[@not_from_here4477]

nailed it!

[@gladiatorsfc9952]

7:35 R Squared Interpretation

[@maxh5522]

Why did I watch an hour long lecture that you covered in ten minutes?!? Our teacher should’ve just sent us to your channel

[@jossri]

Great video!. Thank you. I have a question more related to the type of object of the Lm output. If I’m doing several Lm, how can I extract values of the output and append in a df to compare the results?

[@dataslice]

Great question! You can save the output of the model by assigning the summary of it to a variable, e.g. `x = summary(fish_model)`. Then, if you open up 'x', you'll notice it's a list object with different vectors and values -- you can do `names(x)` to see the different variables and access them accordingly. For instance, if you wanted to extract the r squared, you could call `x$r.squared`

[@jossri]

@@dataslice thank you so much for your answer.

[@justin2icy]

is the "-433.576" known as the regression coefficient?

[@ianleboo]

It is considered the mean of the regression equation when predictor variables are all zero.

[@BarcenasJoel]

How would you interpret the estimate for intercept?

[@dataslice]

Depending on what data you're trying to fit, the interpretation may vary. For instance, if we were fitting a line on a dataset of ages and weights (Y = weight, x = age), the interpretation of the intercept would be 'how much a person weighs when they are born (age 0)'. However, if you're fitting a line to data that's very far from the x-axis, your interpretation may be invalid

[@BarcenasJoel]

What does it mean when you have a std. error that is higher than your coefficient?

[@dataslice]

Again, it's hard to tell without seeing the data but typically this means that the variable may have no statistically significant effect. This could happen because there's actually no effect, or there are some outliers in the data that's affecting the fit

[@nooberinho]

Your video was helpful for some working I'm doing atm. Hopefully you could help me with this question I've asked elsewhere: Hello, I've seen many statistics courses note that for a single linear regression if you regress an outcome variable on a binary predictor variable the slope coefficient is the same as the difference in average outcomes between the two groups. Is this still accurate for a multiple linear regression for a binary predictor variable when you also have multiple other non-binary variables? Thanks!

[@dataslice]

If you have a multiple linear regression with one binary predictor and multiple non-binary predictors, then the slope of the binary predictor is the same as the difference in average outcomes between the two groups *if you hold all other predictor variables constant*. An example of this would be if we were plotting weight as a function of height and gender (weight ~ height + isMale). If the coefficient for isMale is 20, then holding height constant, the difference between the avg male and female is 20. Note that this would be different if there were any interaction effects between isMale and height

[@nooberinho]

@@dataslice that's great, thanks very much

[@noamills1130]

@dataslice What if you have a categorical predictor that has more than two possible values? Would you have to use a different kind of regression model?

[@dataslice]

@@noamills1130 Great question. So going back to the previous example with plotting weight as a function of height and gender (weight ~ height + isMale), let's say we had a 'race' variable that could either be white, hispanic, asian, or african american. We could then create 3 additional dummy variables (isHispanic, isAsian, isAfricanAmer) in our data for the regression. If isHispanic = 1, isAsian = 0, and isAfricanAmer = 0 then that represents the race as hispanic (and so on and so forth for asian and african american). If all three dummy variables are 0, then the person would be white (the baseline). When you make the regression, each dummy variable would be given a coefficient which could help you determine the prediction among the different races. Does that make sense?

[@noamills1130]

@@dataslice Great, thank you so much! I'll be using this for my research project analyzing wildfire trends in the US.

[@nol-sor1985]

10 min >>>>>>>> 2 trimesters

[@dataslice]

Part 2 (Regression *Plots* Explained): kzbin.info/www/bejne/qJeraKN5e9t5ipI

[@doumansarouei4523]

very helpful, thanks