Hi All,
I requires the Data Preparation module, but there is "Identify Unusual
Cases". The command syntax is DETECTANOMALY. I think that Two-Step is
certainly scalable to large data sets. In fact, DETECTANOMALY runs
Two-step first, and then finds cases that are distant to their "peer
group". You could probably do a pretty good job with Two-step and
further analysis with or without the module, but the it is pretty
convenient if you have it.
Hope that helps.
Keith McCormick
www.keithmccormick.com
On 10/11/07, Luca Meyer <
[hidden email]> wrote:
> Is there a method which allows to spot data anomalies (especially
> influential and outliers) in a multivariate context where data are a mix of
> numeric and categorical variables?
>
> I am thinking of hierarchical clustering-like techniques, but here I cannot
> use CLUSTER because I have several thousands cases and the dimension of the
> relative matrix it's just to large to function in such an instance...
>
> Thank you for any suggestion.
>
> Luca
>
> Mr. Luca MEYER
> Market research, data analysis & more
> HYPERLINK "
http://www.lucameyer.com/"www.lucameyer.com - Tel:
> +39.339.495.00.21
>
>
>
> No virus found in this outgoing message.
> Checked by AVG Free Edition.
> Version: 7.5.488 / Virus Database: 269.14.7/1062 - Release Date: 10/10/2007
> 17.11
>
Locked
