HPTMINE multi-term list

JacquesR — Thu, 05 Aug 2021 07:35:20 GMT

Hi all

I am trying to use PROC HPTMINE to parse some text.

I have some multi-term lists that I want to be treated as units, and this page "seems" to explain how it should be done:

https://documentation.sas.com/doc/en/tmhpprcref/14.2/tmhpprcref_hptmine_sect008.htm#tmhpprcref.hptmine.par_mutt

There is also this page, which defines a slightly different format:

https://documentation.sas.com/doc/en/tmref/15.1/n0u9wgweoizcpqn172txiyx8m5h9.htm

This points to the SASHELP.eng_multi dataset, which does seem to follow the latter format.

However, when I run my code, it won't accept a SAS dataset and when I point SAS to a file as the first link suggests, I get an error telling me that the procedure does not recognize my multiword terms, that I should check the multiword format, and that the multiword list will thus be ignored.

I have tried to create a file like this (the headers are commented out because I tried, and failed, with and without the headers):

DATA _NULL_;
    FILE "c:\Users\..\multi_word phrases.txt";
    /*PUT "Term: Token_type: Role";*/
    PUT "Not Recorded: 3: Noun";
    PUT "not recorded: 3: Noun";
    PUT "Potassium Permanganate: 3: Noun";
    PUT "potassium permanganate: 3: Noun";
    PUT "Oxalic Acid: 3: Noun";
    PUT "oxalic acid: 3: Noun";
    PUT "sodium valproate: 3: Noun";
    PUT "small piece: 3: Noun";
    PUT "outer covering: 3: Noun";
    PUT "unknown tablet: 3: Noun";
RUN;

So how exactly should this file be created, because I have tried to follow the hep instructions and just can't figure it out.

I have tried with column-spaced entries, with and without colons, and any other configuration I thought might work.

Thanks

Re: HPTMINE multi-term list

JosvanderVelden — Thu, 05 Aug 2021 16:54:49 GMT

Have you seen the example here: https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/tmhpprcref/tmhpprcref_hptmine_examples07.htm

Which version of miner are you using?
Best regards, Jos

Re: HPTMINE multi-term list

JacquesR — Thu, 05 Aug 2021 22:47:22 GMT

Thanks Jos

Simple as that. No spaces around :3:Noun

Pity that https://documentation.sas.com/doc/en/tmhpprcref/14.2/tmhpprcref_hptmine_sect008.htm#tmhpprcref.hptmine.par_mutt shows the spaces.

topic HPTMINE multi-term list in SAS Procedures

HPTMINE multi-term list

Re: HPTMINE multi-term list

Re: HPTMINE multi-term list