Text Analytics - Is there a way to have one to many words correlation in RapidMiner?

svtorykh

Hi RM support! I know how to have one-to-one word correlations with Correlation Matrix. Is there a way to have words correlated in one to many relationship? For example in my text I can see that Amazon is correlated with Prime. I also know that in the text Amazon, Prime and Shipping are also interchangibly correlated. So how I can show all 3 of them being correlated using RapidMiner?
Thanks much!

svtorykh

Hi Scott,

 

So I have rerun my data through Association Rules. What is the definition of Premises and Conclusion? Couldn't find it in operator description.

Thanks much!

sgenzer

hello @svtorykh - welcome to the community. Seems more like an association mining problem than correlation. Have you tried that approach?

 

Scott

 

svtorykh

Hi Scott! Thanks for your reply. Is it the process utilizing Fp-Growth? I think I have that but results are difficult to interpret in terms of rules generated.

sgenzer

yes exactly. I'd recommend going through a tutorial. Have you looked at the Market Basket Analysis Template?

 

 

Scott

 

 

svtorykh

Hi Scott,

 

So I have rerun my data through Association Rules. What is the definition of Premises and Conclusion? Couldn't find it in operator description.

Thanks much!

lionelderkrikor

Hi @svtorykh

 

Here a description of Premises and Conclusion : 

 

 

Regards, 

 

Lionel

 

 

 

svtorykh

Thank you!