Principal Component Analysis (PCA): With Practical Example in Minitab

[@alsteiner7602]

This is clear, concise, and presented well and in a logical sequence. OUTSTANDING!

[@learnandapply]

Thank you so much for your valuable comments and appreciation 🙏

[@kausalyaakannan7064]

Universities shall pay half of the tuition fees to youtubers for delivering contents with such an awesome explanation😁 Thank you so much sir for this video. How to know sir whether we have to standardize the data based on the output?

[@learnandapply]

Thank you so much for your valuable comments and appreciation! Subject matter expertise is required in that case. If you don't have it, then need to consult with people with related expertise.

[@deepakmoda3401]

Superb way of teaching, Sir!

[@learnandapply]

Thank you for your valuable comments and appreciation! 🙏☺️

[@jensonrozario]

Super informative video. I was looking all over the internet, finally... You did it...

[@learnandapply]

Thank you so much for your valuable comments and appreciation 😊🙏

[@mv829]

The best explanation on KZbin so far, thank you!!

[@learnandapply]

Thank you so much for your valuable comments and appreciation ☺🙏

[@jdo9102]

Vijay, thanks for your invaluable videos. I am a green belt certified now. I am looking forward for more tutorials from you up to the Black Belt .Level. Bless you

[@learnandapply]

Thank you so much for your valuable comments.

[@jayrajjavheri8740]

Fantastic!!.. speechless keep it up! you are serving the people.. god bless you.

[@learnandapply]

Comments like this make my day☺🙏 Thank you so much for your valuable comments and appreciation 🙏☺

[@patrunikiran]

Thank you, indeed good example you have taken for explanation. I am a new learner for PCA

[@learnandapply]

That's great! Thank you so much for your valuable comments and appreciation ☺🙏

[@thaynaalmeida7055]

Thank you for this simple and objective explanation!

[@learnandapply]

You're welcome! Thank you for your valuable comments and appreciation! 🙏☺️

[@marlonmojica7473]

The presentation is well explained. Very helpful to all students and instructors.

[@learnandapply]

Thank you for your valuable comments and appreciation 🙏😊

[@mdmahmudulhasanmiddya9632]

Outstanding performance sir.Ur teaching is adorable sir. Don't say please like. U deserve more than like or subscribe.

[@learnandapply]

Comments like this make my day 🙏🙏☺ Thank you so much for your valuable comments and appreciation 🙏☺

[@terryliu3635]

Wow! One of the best explanations on PCA!!

[@learnandapply]

Thank you so much for your valuable comments and appreciation ☺🙏

[@FaizalKasim_UNG]

Thank you. Your tut is excellent, with clear in steps but concise

[@learnandapply]

Thank you for your valuable comments and appreciation! 🙏☺️

[@learnandapply]

Thank you for your valuable comments and appreciation! 🙏☺️

[@marciabelldbampaha5149]

Good presentation and the baby music is too cute.

[@learnandapply]

Thank you so much for your valuable comments and appreciation 😊🙏

[@ramvemula4336]

Excellent explanation. Thank You very much.

[@learnandapply]

Thank you so much for your valuable comments and appreciation ☺🙏

[@ubhalerao]

Very useful video. My doubts have got cleared.

[@learnandapply]

That's great! Thank you so much for your valuable comments and appreciation!

[@apekshatiwari9290]

Great presentation. Thank you! So we know PC1 is positively correlated with 4 variables and PC2 is negativley correlated with 2 variables. What next? What do we do with this information?

[@learnandapply]

Thank you for your valuable comments and appreciation 🙏😊 Please use this information (PC1 and PC2) to the group variables as per their similarities and you can name them as a meaningful criterion to take a decision. I have explained it in very detail in the video. I will request you to please revisit to understand it in more detail. 😊

[@harasaragajadeera7940]

Very clear and informative. Keep up the good work !!!

[@learnandapply]

Thank you so much for your valuable comments and appreciation 😊🙏

[@wangjessica1275]

How do you explain PC3 ? The third component has large negative associations with income, education, and credit cards, so this component primarily measures the applicant's academic and income qualifications

[@wangjessica1275]

Does it mean increasing income, education and credit card will decrease PC3?

[@learnandapply]

Please look at the contribution of income, education and credit card - it's lower 13%. Need to focus on 1st two components as they are the major contributors.

[@qsdqdqd123]

@@learnandapplyso it means that we need to drop the two original variables (income and education)? you said in the video that sometimes we need more than 90% of the variance explained = 4 principal components. But in the end we only have 2 principal components to analyze the loan applications? I’m quite confused…

[@learnandapply]

It's like a Pareto principle. How much data do you want to consider for taking action?

[@zeeshanazam5104]

very informative, really apperciated

[@learnandapply]

Glad it was helpful! Thank you for your valuable comments and appreciation. 😊🙏 You can also learn it in detail with my mentoring support at vijaysabale.co/multivariate

[@shafiqulislam2663]

I don't know how many thanks should I give you. From last 10 days I have seen more than 15 videos and read many papers.but none of them was easy to understand.thx,thx,thx. Is there any free version of minitab sir? And when to perform PCA and when PCoA sir?

[@learnandapply]

Thank you so much for your valuable comments and appreciation! PCA is used when your are having too many variables and you want to group them logically for easy interpretation.

[@ammabadi747]

Very nice explanation. Thank you very much sir

[@learnandapply]

Thank you so much for your valuable comments and appreciation!

[@priyamishra9893]

Very informative thank u sir

[@learnandapply]

Most welcome! Thank you so much for your valuable comments and appreciation 🙏😊

[@sreelaxmib8941]

Hi, great video, this is extremely helpful. I had a few doubts, 1. In my data set when I do the same, I do not get eigen vector values close to the proportion value. What does that mean? 2. I have another data set with 96 variables. Can i use this method in minitab for this high number of variables? 3. You had said 4 principal components have been chosen, what do you do with the rest of the 3 principal components chosen? Thank you in advance.

[@learnandapply]

Thank you for your valuable comments and appreciation! 🙏☺️ 1. Eigenvalues and proportion are different things, but both are indicators of the contribution of the respective Principal Component. One explains the value, whereas another explains the percentage. 2. Of course, please try it. 3. We are selecting the most contributing Principal Components like Pareto.

[@navadeepkalita456]

Hello. I joined as a member today. Kindly let me know how do we interpret PC2 and PC3 results

[@learnandapply]

Hi Navadeep, thank you for being a part of the community. The principal components mean a category of variables that we are grouping by their similarities. PC2 and PC3 are the 2nd and 3rd groups of variables.

[@saynaislamdibasaynaislamdi8875]

Thank you

[@learnandapply]

You're welcome! Thank you for your valuable comments and appreciation. 🙏😊

[@manzoorahmad-mu3xv]

Fantastic Fantastic

[@learnandapply]

Thank you so much for your valuable comments and appreciation ☺️🙏

[@omerfarukunal110]

Great Presentation, I have a question. My matrix has 162*2076 dimensions. Can I analyze this matric in minitab? How can I do ?

[@learnandapply]

Thank you for your valuable comments and appreciation ☺🙏 Please use Factor Analysis for this analysis. Use the path: Minitab-Stat-Multivariate

[@omerfarukunal110]

@@learnandapply Thank you, matrix is the BOM List ( Products * Materials). So, I'm not sure to use factor analysis. Actually, I want to do k-means but you know see again dimension error :(

[@nabilanursafha]

Im still confused how did u know the variabel correlate with the principal component? It bases on proportion? So the nearst variable to propotion is correlated?

[@learnandapply]

Please check for the highest values of the eigenvectors.

[@uzmafatima2588]

Sir I have indoor air pollution data of 9 pollutants, and questionnaire data of households(socioeconomics, house features and product,health conditions )....how can I use this for my data ....Kindly please guide.

[@learnandapply]

Hi Uzma, you can use both the options Factor Analysis or Principal Components Analysis in this case. If you have some response y, on that you want to see the impact of all these 9 pollutants, then please use regression analysis in that case.

[@yamikanikaliwo2084]

well explained big up buddy

[@learnandapply]

Thank you so much for your valuable comments and appreciation! 🙏☺️

[@NicholeRojas-r8i]

Hello! what criteria do you use to eliminate outliers?

[@learnandapply]

This is based on Mahalanobis distance. The Mahalanobis distance measures the distance from each point in multivariate space to the overall mean or centroid, utilizing the covariance structure of the data.

[@mdmahmudulhasanmiddya9632]

Sir in .41mint in this vedio shows the correlation between original variable and PCA component.or it is different thing.please reply sir.

[@learnandapply]

We are grouping variables as per their similarities for easy understanding and interpretation of results. This grouping is called as Principal Components.

[@mdmahmudulhasanmiddya9632]

@@learnandapply this is Eigen vectors

[@learnandapply]

Yes, this is weighted and grouped based on eigenvalues 👍

[@mdmahmudulhasanmiddya9632]

@@learnandapply thank u sir

[@nathalieramos5942]

Thank you very much I am from Peru and it helped me a lot. I just have one question, how can I assign weights to my variables. Can the highest results for component one be the weights for my financial stability indicators? and if so, what would I do with the values that come out with a negative sign? Thank you so much for everything.

[@learnandapply]

Thank you so much for your valuable comments and appreciation 🙏☺. Yes, absolutely. The components having higher eigen values need to be select first. The negative sign indicates that it is impacting adversely.

[@nathalieramos5942]

@@learnandapply THANKS FOR YOUR HELP :)

[@learnandapply]

You're welcome and thank you for your valuable comments ☺🙏

[@yagusti_n]

Sir....how to analyze principal components manually, and how to get eigenvector values ​​by manual calculation

[@learnandapply]

That's a great question. We can calculate them by using formulae for eigenvectors and eigenvalues. I think I should create a video on this topic.

[@yagusti_n]

@@learnandapply I will wait for the video sir. thanks very much.

[@yagusti_n]

@@learnandapply sir.... make a video of the manual calculation of the principal component analysis (eigenvector and eigenvalue) using the data in this video. Thank You so much Sir...

[@learnandapply]

I think this video can help you to know how it is coming: kzbin.info/www/bejne/jqe5aHStgatrrJY

[@capecoaster69]

good explanation !

[@learnandapply]

Thank you so much for your valuable comments and appreciation!

[@conanannisa1811]

Terimakasih atas penjelasannya, sangat membantu

[@learnandapply]

Thank you so much for your valuable comments.

[@cyrilsantos3610]

Sir, how did you know what each component measures? You said that age, residence, employ, and savings have large positive loadings on component 1 so this component measures long term financial stability. How did you arrive at long financial stability? Thanks :-)

[@learnandapply]

In Principal components, we need to look at the variables having higher eigenvector values. So, if you look at the first principal component, the variables you had mentioned are having higher eigenvectors. Now, how to name them? Well, you must be a subject matter expert in that field. If not, you need to take a help from subject matter experts 😊

[@cyrilsantos3610]

@@learnandapply Many thanks Sir :-)

[@learnandapply]

Your welcome 😊

[@andpinto1]

Your answer is sound, because eigenvectors only tell you how relevant the variables in the overall variance. So what you do is checking which have the highest eigenvectors and go check independently what they correlate to. This depends on your expertise. Here, the Manager would see he/she would have to look into residence, employment, age and savings. That´s as far as PCA goes. It also tells you which samples are more similar, ie cluster together.

[@localguy123]

I have a large Dataset consisting of two variables, Voltage and Time. Can I do PCA on it?

[@localguy123]

And can we do curvilinear component analysis?

[@learnandapply]

For 2 variables with large data set, you won't need to go for PCA. Just use a regression model. If you want it in more detail, please use nonlinear regression.

[@sonamchavan9346]

Can you please explain the MNIST Handwritten Digits with PCA

[@learnandapply]

Can you please elaborate on your question? Thank you.

[@abhijeetdas6279]

How can I export the output graphs from Minitab?

[@learnandapply]

Just right-click on the graph and export it to Word or PowerPoint.

[@sankar_riser]

Sir I've a doubt can I use replicated data for one variable

[@learnandapply]

Thank you for your valuable comments. Try not to use replicated data. It will create an error.

[@Shabbir2749]

Nice work bro. Please make video on GLM

[@learnandapply]

Thank you so much for your valuable comments. Sure, I will do it in future videos.

[@Shabbir2749]

@@learnandapply thanks all the best

[@pravakirandash9758]

Thanks for the video,Sir..How can I install minitab software? Is it free.

[@learnandapply]

Please check my video on statistical software to get all the details. This is a 30-Days FREE trial. Anything else that I can help with?

[@pravakirandash9758]

@@learnandapply I will check, Sir.. Thanks a lot for your help.

[@razheer100]

My issue is that I downloaded minitab express per my Universities free trial. Yet under stats, no multivariate option is available to do a PCA. Any suggestions?

[@learnandapply]

Please try reinstalling it, otherwise, this software is with some fewer options.

[@emiliooo2877]

te amo

[@recepgunyuz3121]

where is the scores ?

[@learnandapply]

These are eigenvector values. You can get it from PCA output table.

[@DevanshiHingrajiya]

can you please share the data

[@learnandapply]

Thank you for your interest in learning this important topic and your valuable comments. For in-detail learning of this topic with data, notes, videos, and my mentoring support, please visit - vijaysabale.co/multivariate

[@ytubeleo]

The same Christmas song again on repeat?! Other than this it was very good.

[@learnandapply]

Thank you for your valuable comments. This is a video uploaded a year before. Sorry for the inconvenience caused to you. 🙏🙏