Interpreting the Summary table from OLS Statsmodels | Linear Regression

In this video, we will go over the regression result displayed by the statsmodels API, OLS function. We will go over R squared, Adjusted R-squared, F-statistic & T-test of the feature values.
Link to the notebook : github.com/bha...
If you do have any questions with what we covered in this video then feel free to ask in the comment section below & I'll do my best to answer those.
If you enjoy these tutorials & would like to support them then the easiest way is to simply like the video & give it a thumbs up & also it's a huge help to share these videos with anyone who you think would find them useful.
Please consider clicking the SUBSCRIBE button to be notified for future videos & thank you all for watching the video.
You can find me on:
Blog - bhattbhavesh91....
Twitter - / _bhaveshbhatt
GitHub - github.com/bha...
Medium - / bhattbhavesh91
#LinearRegression #statsmodels #interpretation

[@joemireku7104]

I am glad you created this video 4 years ago.

[@bhattbhavesh91]

Glad it was helpful!

[@suhailchougle7315]

It's like the concepts which I couldn't find anywhere else on youtube, I end up here. This is exaclty what I was looking for, the explanation of this "ols.summary()".I believed I wouldn't find any. But here I am. I really really appreciate this buddy .Thank you so much.

[@bhattbhavesh91]

Glad it was helpful!

[@dhananjaykansal8097]

This is just too good Bhavesh bhai. Please keep this work going. It's so helpful that I can't put it into words. Thanks a lot.

[@jorgemarin7]

Thank you Bhavesh!! Just what I was looking for and very well explained.

[@bhattbhavesh91]

Glad you liked it

[@WillC]

Brilliant! Thank you Bhavesh.

[@brunocoelho9088]

I don't think you should get into the habit of filtering warnings - this may hide important information - The correct way is to only hide the specific type of warnings you want to.

[@asadtanvir4065]

Hi Bhavesh, thank you for this nice interpreting video. I have two questions though: #1 You didn't go through the Df residuals, Df model, Log-likelihood etc. stuffs. It would be good if you could cover those in this video, OR at least state the significance of those parameters here in the comment section. #2 Is there any way to get SSE (sum of squared errors), SSR(sum of squared regression) values of this fitted model?

[@bhattbhavesh91]

I wanted to keep the video simple so skipped some parts! I'll cover the remaining topics in the near future videos! Thanks

[@souhamahmoudi7745]

Thank you very much for this tutorial, it's really hepful !

[@ruteshrathod3844]

thanks, Bhavesh Sir, for making concept clear about feature selection

[@Sharukkhan-kx9rw]

Thanks for the Great Video, Could you tell me What the number represents [0.025 0.975] ?

[@bhattbhavesh91]

I'll create a new video around this soon!

[@areejawanareejawan1615]

It's really amazing thanks for such informative and highly understandable videos.

[@prabhakaranshanmugam5295]

Please Explain about omnibus, durbin watson jarque bera etc

[@bhattbhavesh91]

I will cover this in the next set of videos!

[@SaffronKage]

Super thanks such a simple and accurate explanation being from programming background the stats summary interpretation was bugging me a lot

[@bhattbhavesh91]

Glad you liked it!

[@SheikhMozaffarHossain]

How does OLS ewquation look like?

[@abhijeetpatil6634]

Thanks Bhavesh for such a beautiful video

[@kancharlasrimannarayana7068]

we have to remove outliers and do the ols model?

[@julesdrums6167]

Thanks! Extremely helpful. What about log-likelihood, AIC, and BIC?

[@bhattbhavesh91]

Will upload soon

[@lokeshrathi5500]

You making these videos as early as 6:47 AM shows your hard work and passion for this! Thank you so much for your Work! Question:" The condition number is large, 2.9e+04. This might indicate that there are strong multicollinearity or other numerical problems." - what do you infer from this?

[@GauravSharma-ui4yd]

Great Explanation :). But you skipped some parts. I also want to know what is AIC, BIC at upper right. And what is [0.025 0.975] columns after the t-test columns. What is that omnibus? Are the skew & kurtosis are of response distribution? And all the remaining statistics at the bottom.

[@bhattbhavesh91]

I wanted to keep the video simple so skipped some parts! I'll cover the remaining topics in the near future videos! Thanks

[@GauravSharma-ui4yd]

@@bhattbhavesh91 thankyou for the reply and please cover that asap.

[@bhattbhavesh91]

@@GauravSharma-ui4yd - I will try! I'm loaded with work at this point of time!

[@shubhamkumarjha1868]

@@bhattbhavesh91 Can you share the link if you have posted the video

[@PavanKumar-uw7si]

Great Explanation dude!!

[@victorperezandres7966]

Amazing video. Really helpful. Simple & clear. Thanks a lot!

[@bhattbhavesh91]

You're welcome!

[@faiquejatu8649]

hellow sir, one thing i would like to ask is what optimization algorithm does stats.model.api OLS uses, ? just like sklearn linearRegression uses ClosedForm solutions. Thank you in advance 😊

[@vipultyagi74]

It uses Gradient descent

[@SumitKumar-uq3dg]

Hi Bhavesh. Really a big fan of ur videos. U have made all topics so easy to understand. Can u pls make a video on evaluating a logistic regression model with statistics like ks test, psi, concordance pls

[@bhattbhavesh91]

Thanks for your kind words Sumit! I'll make a video on how you can evaluate logistic regression soon!

[@anjan700]

@@bhattbhavesh91 Hi! Is the video published? If yes can you post the link here.

[@tejasbiraris9077]

This was quite helpful. Thankyou so much

[@bhattbhavesh91]

Glad it was helpful!

[@389_kavetiupender2]

Hello sir, I'm working on a project, I have a dataset less 100 values. I have created a regression model but I am ending with huge MSE value. Sir can u suggest any idea or techniques so that my model performance will be improved.

[@ice-skully2651]

Thank you for the explanation, though I have an issue with understanding why we should omit feature 3, it is the only feature with a high p-value and therefore fails to reject the null coef = 0 hypothesis meaning that there is a linear relationship between it and the target.

[@I3oblivion]

Hi, I have a question about my specific results. Would I be able to send you my statsmodel summary results and you can help me interpret them for my case?

[@RajibDas-kq2uz]

if we have categorical and continuous combine my X and continuous Y , i have categorical variables and my target is continuous variable. that time what we will do ? i can check p value like this for categorical variables is well

[@fahimmuntasirniloy]

Thanks for the quick and concise explanation.

[@heitorosorio1847]

Thanks for the video. What about the column "Std err" ?

[@daveneifer2456]

Thanks so much for the breakdown of the results table, it was very helpful.

[@bhattbhavesh91]

You're very welcome Dave!

[@hemchandrapatil8697]

If I have a categorical variable, then how to identify that it is significant or not?

[@fvc1612]

Are those the results of causality or only a significant relation between variables?

[@manishkumarsinghwxkvdgjjix3139]

Sir its really helpful. But i have a as if R squared value is far for 1 and p>t is also less then 0.05 then what we do ??

[@SHUBHAMKUMAR-ty6bh]

Bhavesh you are great. Very nice interpretation.

[@ZhiboXiao]

The clearest explanation! Thank you!

[@bhattbhavesh91]

Glad you think so!

[@binyameenkv5068]

Thank you so much... That was a really informative video! I just couldn't understand one thing, why are we adding a constant column and appending it to actual x column?

[@chiragagrawal7104]

how to improve the condition number??

[@md.sultanmahmud1321]

Hi Bhavesh, Can you please tell me about the beta coefficient? Which one is the beta coefficient in the summary results?

[@sriram_cyber5696]

Amazing Tutorial

[@bhattbhavesh91]

Glad you liked it

[@preeethan]

Can we use VIF also to evaluate the features.?

[@hemangdhanani9434]

awesome explanation

[@bhattbhavesh91]

Glad you liked it

[@Mrjacksurvival]

Great video. How to put uncertainty of measures in the regression?

[@anasali4782]

Great video!

[@bhattbhavesh91]

Glad you enjoyed it

[@ncatch25]

Could you please explain how the Durbin-Watson output is interpreted?

[@bhattbhavesh91]

I have created a video on Durbin Watson test! Do have a look - kzbin.info/www/bejne/fJqlc6OpmMdpm9U

[@robertomatsumotocobra7684]

Thank you very much, it was really elucidative!

[@bhattbhavesh91]

Glad it was helpful!

[@mobbie16995]

Thanks great explanation - you're the best!

[@bhattbhavesh91]

Glad you think so!

[@joyvictor6165]

This was absolutely useful. Thank you

[@bhattbhavesh91]

You are welcome

[@kailashsinghrautela8253]

Nice and concise thanks very much

[@bhattbhavesh91]

Glad it helped

[@dcshortshorty475]

DEEPESH CHADHA

[@AmandeepKaur-ot1tn]

what is the method to save stats model sir.

[@bhattbhavesh91]

I didn't fully understand your question!

[@ronnawatprangtip8555]

Thank you for your clear explain:>)

[@bhattbhavesh91]

Glad it was helpful!

[@nitul_singha]

thank you

[@daphnes1001]

thank you so much!

[@antonyvibin3274]

Good one buddy :)

[@siddharth4251]

nice explanation brother!

[@bhattbhavesh91]

Appreciate it!