How to generate recommendation list based on service notes prepared by customer support

As part of my PoC, I should generate recommendation list based on service notes prepared by customer support of a start hotel say XYZ.

The notes are as below
Name_Id: 10306 on different annexed hotels say,
10306    Tennis / Daughter 18 YO plays in tournaments, Son 8 YO
10306    Prefers travel, local food & drinks  Dec 2012 Room was not ready upon arrival
10306    L'Occitane products / Shower Gel Daily / extra hand towels

Name_Id: 43736 on different annexed hotels say,
43736    Occasions - Anniversary Wedding/Nov 11, 1999.
43736    Spa, Extra Pillows 
43736    Mr likes French, Italian, Mediterranean and Seafood Cuisine

From above notes, we prefer to generate preferences grouped by Name_ID on various hotels say,
10306    GF, WS, LP, ELC 
43736    SP, KB, PF

Note: GF(Golf), WS(Water Sky), LP(L'Occitane products), ELC(Exploring Local Culture), KB(King size bed), PF(Pillow foam), SP(Spa)

Can someone guide me how to proceed and the best node that has to be chosen?


Hi Amalan,

one approach would be to search in the grouped notes for terms with a term or wildcard dictionary and replace the terms by the markers (GF, WS, LP, ...). You would need to transform the grouped (concatenated) strings into documents and use the dictionary or wildcard tagger. Of course you need a dictionary for that. With the replacer node it is possible to replace strings, that match on specified regular expressions with another string.

Cheers, Kilian

@Kilian Thanks. I am on same approach. Since am new to this tool I find few basic complexities. I shall spend sometime on this to get some decent queries.