topic Re: Is it possible to do a NLP like extraction using Data Integration tool in Talend Studio

Is it possible to do a NLP like extraction using Data Integration tool

Anonymous — Thu, 18 May 2017 16:51:03 GMT

Hi All,

I want to know whether NLP like extraction is possible with Talend Data Integration tool.

My requirement is as follows:

In an excel sheet ,I have a column named 'Description'. It is a lengthy string giving details about a product.

I have a column in DB,named 'Product Type'. Product Type can have values like Product1,Product2,Product3 etc.

I need to analyse the 'Description' field in the excel sheet and find out whether a value like Product1 or Product2 or Product3 is available in that .If its present,extract the value and populate the 'Product Type' column in DB.

Please help me to figure out the feasibility of doing such a task using Talend DI.

Thanks,

Sreedevi

Re: Is it possible to do a NLP like extraction using Data Integration tool

Anonymous — Thu, 18 May 2017 17:01:10 GMT

of corse, u can do it.

u could use the equals or conten Equals function to realize this.

Re: Is it possible to do a NLP like extraction using Data Integration tool

TRF — Thu, 18 May 2017 17:16:13 GMT

IMHO, using just "equals" or "contains" methods may be a little restrictive compared from what you can expect from a NLP method.

I suggest you to include Stanford CoreNLP (or other NLP engine) in your project if you want to take advantage of a real natural language analysis tool.

Re: Is it possible to do a NLP like extraction using Data Integration tool

Anonymous — Fri, 19 May 2017 07:00:45 GMT

Thanks all.

Sreedevi

Re: Is it possible to do a NLP like extraction using Data Integration tool

TRF — Fri, 19 May 2017 22:53:15 GMT

@sreedevi_s, let us know if it helps and what's the next step for your usecase

Re: Is it possible to do a NLP like extraction using Data Integration tool

Anonymous — Tue, 23 May 2017 18:32:41 GMT

Hi,

Yes.It helped to get an idea about the steps to follow.I created a lookup table with possible values and merged it in tmap.

Thanks,

Sreedevi