Your data analysis solution
Name is required.
Email address is required.
Invalid email address
Answer is required.
Exceeding max length of 5KB

PCA Analysis for spectral Data ( absorbance Vs wavelength)

By Steph | Oct 23, 2017 11:41AM CEST

Hi, I've only just started using XLSTAT. I'm currently trying to run PCA analysis on a large data set (absorbance Vs wavelength). The pre-treatment I would apply is mean centring (but i see that XLSTAT does that anyway). I have ran the analysis 9 times and each time the output is wrong!!!! I have watched ALL the tutorial videos and I am following instructions accordingly.

EVERYTIME i run the PCA analysis;

I keep getting the following message;
''The Bartlett's sphericity test cannot be computed because of a multicolinearity between the selected variables. You can use the multicolinearity analysis tool to detect and remove the variables responsible for the problem''

Also, for my Scree plot, ALL the points are located outside of the graph!! and this is also the case for my 'relationship between variables graph'.

Please help... What am i doing wrong ??
Do you use PCA the same way for spectral data or should i be using something else????

I have been working on this for hours. Nothing works :)

How do you preform PCA on a LARGE Spectral data-set (I have 22 samples and 175 absorbance's) that need to computed against a variet of wavelengths.

How do you so this ???
Anyone?

Thank you so much for your time.




Best Answer
By Sébastien | Oct 23, 2017 12:36PM CEST | XLSTAT Agent

Hi Steph,

PCA can definitely be applied on your data.
However, you probably must have previously activated some undesirable options that make it not work properly. For instance, Bartlett’s sphericity test is likely to detect multicolinearity among 175 variables, even by chance.
I would suggest that you reset the dialog box by clicking on the round blue arrow at the bottom left, and try configuring things again. Make sure you don’t activate Bartlett’s sphericity in the Outputs tab.
Let us know if you keep obtaining weird results.

Jean Paul


Up

0

Down

By Sébastien | Oct 23, 2017 12:36PM CEST | XLSTAT Agent

Hi Steph,

PCA can definitely be applied on your data.
However, you probably must have previously activated some undesirable options that make it not work properly. For instance, Bartlett’s sphericity test is likely to detect multicolinearity among 175 variables, even by chance.
I would suggest that you reset the dialog box by clicking on the round blue arrow at the bottom left, and try configuring things again. Make sure you don’t activate Bartlett’s sphericity in the Outputs tab.
Let us know if you keep obtaining weird results.

Jean Paul

Post Your Public Answer

Contact Us

1c26995d494fb3061dd0ae8571ffc0a4@xlstat.desk-mail.com
https://cdn.desk.com/
false
desk
Loading
seconds ago
a minute ago
minutes ago
an hour ago
hours ago
a day ago
days ago
about
false
Invalid characters found
/customer/portal/articles/autocomplete
9202