Principal Component Analysis in SPSS

[@thecopy9975]

This channel is awesome.. thanks for the great video..

[@theoutlier7395]

Thank you for your support, All the best...

[@PriyaKumari-gx9gt]

how to determine sum of squares (eigenvalues) and percentage of error for given factor scores?

[@theoutlier7395]

spss displays eigen values by default

[@Aarreerreerr]

Sir, how did you made the heat map ?

[@theoutlier7395]

Excel

[@anshul-katyal]

When we have dependent variable so Do we have to select dependent variable also for PCA? And we get factor after this what can we do for the interpretation?

[@theoutlier7395]

use only independent variables for factor analysis

[@deepkhatu1289]

Can you help me Integratinh fuzzy Values & can i use this in Computing most critical safety risk factors

[@khushboo4165]

7:35 Why did you select the absolute value as 0.6?

[@theoutlier7395]

if the factor loading is lesser than 0.6, we neglect it.

[@bhargavireddy5730]

How to make a graphical representation of PCA?

[@theoutlier7395]

In SPSS it isnt possible, in statcraft you will get it by default. You could try using Powerpoint

[@pujadey8096]

How did you make the heat map here?

[@theoutlier7395]

in PCA you get the descrive statistics option

[@bianca4397]

sir, my research is similar to yours. i am researching about sensory test for 30 panelists and there is 5 samples. i followed literally all of your tutorial including make the panelists is the variables. but why it said that my correlation matrix "non-positive defini te matrix"?

[@theoutlier7395]

looks like your sample size is very small. at least 100?

[@theoutlier7395]

typically 1:20, for every variable 20 records

[@Valpuia_Sêktak]

Can PCA tells us which original variables to retain for certain statistical analysis?

[@theoutlier7395]

If the loading is greater than .7 retain it

[@neelamkhanoda6252]

Sir in my study have non metric data, I want to apply multivariate analysis, is it PCM method appropriate for it. 1. If not what method appropriate for it. 2. I have socio and economic section related many variables.

[@theoutlier7395]

If you have Ordinal(5 point scale) you can still go ahead and use PCA

[@theoutlier7395]

Pl. Note: Since you have non-metric data, Principal Component Analysis (PCA) might not be the best method. PCA is typically used for metric data (interval or ratio) because it relies on variances and covariances, which aren't meaningful for categorical data.

[@theoutlier7395]

Try Multiple Correspondence Analysis (MCA) or f actor Analysis for Mixed Data (FAMD)

[@neelamkhanoda6252]

Thank you so much sir for your response 🙏🙏

[@theoutlier7395]

@@neelamkhanoda6252 All the best

[@failedupscaspirant7938]

Sir I want to make a vulnerability Index. can you guide pls

[@theoutlier7395]

Hi i am not sure what do you mean by vulnerability index

[@failedupscaspirant7938]

@@theoutlier7395 brief explanation - safety of beach from croc ATTACK IS DEPEND UPON 5 FACTORS. EACH FACTORS IMPORTANCE AND CONTRIBUTION VARIES. SO GY GIVINg a weightage to each factor I want to create an index score.

[@jeronimocliff2972]

Hello my friend. I checked and made the Factor Analysis just like you, but within Excel there is alway just one Sheet. Can you help me so that I can see Heat Map / Eigen Value? I checked the descriptive statistic option in PCA but could not find it directly

[@theoutlier7395]

Hi, i have created the factor loading and the other output by running analysis in SPSS. it is not part of the raw data provided

[@jeronimocliff2972]

@@theoutlier7395 thank you very much for the clarification

[@theoutlier7395]

All the best..

[@moyofoluwaogunyemi2315]

Thank you for this. My result is not displaying the KMO and bartlett result. Could there be any reason for this?

[@theoutlier7395]

if you are using SPSS, pleasse go to analyze menu, dimension reduction, factor, this open up factor analysis dialog box. one you are inside factor analysis dialog box at the right hand side you have descriptive statistics tab. please click on descriptive statistics tab. this will open up a new dialog box . the last option here is KMO and bartketts test of shericity. please make sure you select this click ok. you should get the kmo result

[@theoutlier7395]

if you are having issues please let me know

[@bianca4397]

@@theoutlier7395 i double checked about the step you explained but still dont find any KMO and bartlett result. can it be analyze by excel or anything else?

[@theoutlier7395]

@@bianca4397 SPSS only

[@the-nutriguide-e2b]

i did survey, and had 4 study groups. in that case, is it possible to run PCA on SPSS

[@theoutlier7395]

Yes you can go ahead with PCA

[@the-nutriguide-e2b]

@@theoutlier7395 Could you help me with running PCA in SPSS? I'm analyzing a dataset with 4 study groups and 400 samples to find where risk factors cluster the most. Your guidance would be valuable!

[@priyagupta558]

Sir, can we use this factor as a variable for regression. Also, do we need to consider them as negative values or absolute values?

[@theoutlier7395]

you can use the factor scores as variables

[@theoutlier7395]

Also, When interpreting factor scores from factor analysis, it is important to use the scores as they are, including both negative and positive values. These scores indicate the relative positioning of observations along the factor dimensions. Converting them to absolute values would distort the interpretation and relationships identified by the factor analysis. Negative values are meaningful and represent observations that are below the mean of the factor, while positive values are above the mean.