Text Mining with supervised learning and training data set

steffen_sachsen

Dear all,

i am new to this community, Text Mining and RapidMiner and have a question about whether using RapidMiner is appropriate to my task at all.

I have a set of 24 companies and i read the annual reports (of only one year) to identify specific keywords regarding digital transformation/digitalisation. I then extracted all sentences containing these keywords, the finding page and the finding place within the annual reports and defined specific categories for every extracted term (manually) (i.e. category=product info, strategy, personell etc.). 

Now i want to "train" RapidMiner with this training data set, to identify and extract the keywords in all annual reports form previous years and, if possible, let RapidMiner predict the category based on the input form the training data set. 
In the end i would like to calculate some sort of a scoring model, which company publishes most regarding digitalisation. 

So this would be the task, and my question is: is RapidMiner appropriate for this?

All the annual reports are pdf files, my training data set is a excel file. I already tried a little bit using the text mining extension and read some threads in this community, but i can´t help to find a way to get my task done, especially since i am not good at all with new software.

Thank you very much for reading my post and thank you in advance for any advice!

Best regards

Telcontar120

You should look at the LDA topic modeling operator, which is available in the operator toolbox extension.  

steffen_sachsen

Thank you very much for your help! I will have a closer look at it and hope to get it adopted for my needs.

stefan_koerling

Do you got a solution for the task? I have the same issue.

MartinLiebig

Hi @stefan_koerling ,

did you try LDA?

BR,

Martin

Nadine_12

Hi @stefan_koerling

did you get to a solution for this text mining issue with LDA?

BR Nadine