'nominal correllation matrix' from example set

imke

Hello,

I have already done a text mining process and now I have the Example Set (Process Documents from Data) table. With this I want to calculate how often two words occur in the same text. At first I thought I could use Correlation Matrix Operator, but that does not work. So I tryed with Auto Model the Clustring, but for this I can only take two entrys of the example Set and I want to know it from all the words. So I thought maybe I could add the x-Means Operator in my process, but for x-Means my Data Set is a way to big and with k-Means I'm not getting the results I want. (No Correlation Matrix like with Auto Model anymore).

So my question is: Is there a possibility to create a correlation Matrix with the ExampleSet?

Thank you

Imke

MartinLiebig

Hi@imke,

it feels to me like this is a case for FP-Growth or for n_grams? See attached example.

BR,

Martin

< ?xml更小ion="1.0" encoding="UTF-8"?>



<宏/ >

<运营商激活= " true "类= compati“过程”bility="9.0.002" expanded="true" name="Process">


















Binary term occs

imke

Hello Martin,

that's quite good, but not the right solution for me I think. N-grams are only words which are following themselfs and I want to know, which words are in wich text together, but not directly after the other word. Do you know what I mean?

Greatings

Imke

imke

Hello Martin,

I need to correct myself. With the right settings FP-Growth is perfect for me!

谢谢!

Imke