Introduction to MultiNomial Logistic Regression (Outcome more than two class) & Solution Approach

Get Crystal clear understanding of Multinomial Logistic Regression.
To know step by step credit scoring, model design, multi collinearity treatment, variable selection, model validation etc, please refer to
-----------------
www.udemy.com/logistic-regres...
-------------------------------------

[@nikolaybogdanov5258]

Thank you very much for your outstanding explanation, sir!

[@rajpaul1501]

Very nice explanation. Crystal clear. Cant Thank you enough

[@VISWESWARAN1998]

Thank you for an excellent explanation!

[@lythien390]

Thanks for the video! Very well explained!

[@samarendrapradhan5067]

Very easy and real time example and understanding..Thanks sir

[@markjohnson6537]

Thank you Sir, excellent, clear explanation.

[@finkelstinlu6580]

Very good explanation, thank you.

[@panpeter7879]

A real time-saver!!! Thank you very much!

[@soggypancakes666]

Thanks for the video, Gopal. This approach seems like it only works if the classes in the dependent variable are mutually exclusive. What approach would you use if the classes in the dependent variable are not mutually exclusive? For instance, in the ice cream example, what if the people were allowed to choose one scoop of vanilla and one scoop of chocolate?

[@mohammadbasheer6192]

very good explanation sir.. thanks

[@brunovargassanmartin3660]

ur good teaching, ty

[@leahcim87]

Well explained Gopal! Thanks!

[@21LeonidasZ]

Hello, I would like to ask whether it would be preferable, if possible, to divide a categorical variables into many binary variables and model separate logistic regressions , or run one multinomial logistic regression. I ask because in the former case one could model IVs in the way he wishes to for every outcome, while in the latter all IVs are identical for every outcome. Thank you.

[@winterwilderness5073]

So clearly explained! Thank you.

[@shambhudeo5668]

clearly explained.

[@nortongartino4602]

Thank you for the video!

[@Rusikulya]

Thank you Sir

[@ShameenaKhatoon-qb9yp]

spent more than 4 hours to understand this until i get your video and learnt in 10 minutes.

[@gopalprasadmalakar12]

Thanks a lot Shameena for your feedback !

[@Hemant_2487]

Excellent way of teaching, I loved it. Thank You, sir.

[@fabianwinston7371]

A tip : watch series on flixzone. Me and my gf have been using them for watching loads of movies recently.

[@kaysonvictor180]

@Fabian Winston definitely, I have been watching on Flixzone for years myself :D

[@zafariqbal4537]

Good work, thanks

[@rahiqueabdul392]

U saved me in the test

[@PasaulioValdovas]

Well explained

[@adityanjsg99]

Thank you sir..

[@MikeeBPacot]

This is what i need! Thank you sir

[@xingyuanchen1796]

this video really helps!

[@PavanKumar-ef1yy]

Thankyou

[@nikkigupta4091]

Perfect

[@kartikeyachaudhary4040]

Thank you :)

[@hemanthpeddi4129]

Can we use Gradient descent logistic regression for the multinomial classfication problems??

[@shrgai8821]

Yes

[@inlawbestltdcompany4253]

Thank for this piece. Please help me with this scenario I have three outcomes (States) - State 1.) Not delisted, State 2.) and State ST 3.) Delisted and I want to construct a Multinomial logit model to predict the probability that a firm will go into any of these 3 States. First question, can I call these 3 states as 'unordered'? normally a firm goes from State 1 to 2 then to 3 however, it is also possible for a firm to go from State 1 to 3 without passing through state 2. Secondly, Firms can go from one state to another within a sample period of say 4years? E.g if a firm goes from State 1 to State 3 within a sample period, what state do I consider this firm to be? Thirdly, following on from question 2, please advise how I can gather data for firms in different states without violating MLR assumptions

[@gopalprasadmalakar12]

Inlawbestltd, Here is my answer. First question, can I call these 3 states as 'unordered'? normally a firm goes from State 1 to 2 then to 3 however, it is also possible for a firm to go from State 1 to 3 without passing through state 2. --> Yes Secondly, Firms can go from one state to another within a sample period of say 4years? E.g if a firm goes from State 1 to State 3 within a sample period, what state do I consider this firm to be? Thirdly, following on from question 2, please advise how I can gather data for firms in different states without violating MLR assumptions --> The outcome doesn't seem strictly ordinal as it is not like very satisfied -- satisfied kind of situation. I will suggest you Multinomial Regarssion

[@lewimuchiri]

Looks like a time-indexed kind of a problem. You may use pooled logistic regression (multinomial) as the observations are being made through a time window and each firm has a state at each period of observation.

[@oludipe737]

I wish you can show the formula in an application. This is so confusing though, but thanks for your time.

[@bawashah6730]

can there be just one independent variable?

[@2110babli]

Will the result be same or different for linear and multinomial logistic regression? Can you please tell me the difference, if there?

[@gopalprasadmalakar12]

The dependent variable is nominal (not numeric , not even ordinal), that's why you can't apply linear regression here by any chance

[@treyhannam3806]

Well done

[@simranjitsingh7227]

Sir, what if there rae more than three classes then also multinational regression ?

[@gopalprasadmalakar12]

Yes

[@pherriejamesbacolod6789]

can you share the assumptions of Multinomial Logistic Regression

[@arellinaresh5646]

Sir ur presentation is speechless it helps me a lot thanq sir

[@Gamabunta24345]

so do binomial regression for each outcome then use bays theorem to find the probabilities

[@toonepali9814]

or logistic regression using sigmoid function

[@durgasthan12]

Is Multinomial logistics regression same as Mixed logit? can you help me with this Sir?

[@souheylaghebghoub8223]

mixed logistic means both fixed and random effects are considered

[@Luciiffeerr]

I'm having difficulties to understand WHAT logistic regression is. Is it ex: taking to variables and putting them together so we can se what the probability is between these two in a situation?

[@gopalprasadmalakar12]

This should help kzbin.info/www/bejne/rWGwon95fc2FlcU

[@tubeyou494]

Sorry..I didn’t understood the reference model approach..how come p(a)/p(c) & p(b)/p(c) ?? And what is dependent variable in each reference model?

[@jabulanimolefe9395]

if i have 4 catergories it means i will develop 3 equations right?

[@gopalprasadmalakar12]

Yes.

[@lighthomefood3812]

Thank you sir.....Sir could you please explain univariate, multivariate, binary logistic regression?

[@gopalprasadmalakar12]

Dependent variable can be binary or may have multiple values (but not ordered - hence multinomial) or multiple values but ordered (like very satisfied, satisfied, neutral etc). kzbin.info/www/bejne/rWGwon95fc2FlcU I will suggest to look at above.

[@lighthomefood3812]

@@gopalprasadmalakar12 thank you so much sir

[@toobabukhari7206]

what to do if there is order in the outcome and it is polytomus. Independent variable also has 2-3 categories.

[@gopalprasadmalakar12]

Tooba, I have put a video for ordinal regression as well. Look at that. It will help you. Regards

[@toobabukhari7206]

@@gopalprasadmalakar12 I'll check that. Thank you so much

[@umbnmbu4679]

explain you

[@heyjulyabas1681]

Thank you so much.