topic Re: COLLABORATIVE-FILTERING SYSTEM in SAS Programming

COLLABORATIVE-FILTERING SYSTEM

EloarL — Thu, 29 Aug 2019 12:44:11 GMT

%MACRO PREDICTSCORE(USRCNT=15);
%DO I = 1 %TO &USRCNT;
PROC SQL;
CREATE TABLE USERUSERPREDICTION&I AS
SELECT BUSINESS_ID, USER_ID, &&RU&I AS RU,
"&&USER&I" AS USER, CASE USER_ID
%DO J = 1 %TO &&NCNT_&I;
WHEN "&&&&NNAME&I._&J." THEN
&&&&NEIGHBOR&I._&J. * (STARS - &&RN_&I._&J)
%END;
END AS WEIGHTEDSCORE,
CASE USER_ID
%DO J=1 %TO &&NCNT_&I;
WHEN "&&&&NNAME&I._&J." THEN &&NEIGHBOR&I._&J.
%END;
END AS W
FROM YELP.USER_PROFILE_V1 WHERE
USER_ID IN (
%DO J = 1 %TO &&NCNT_&I;
"&&&NNAME&I._&J" %IF &J<&USRCNT
%THEN ,;
%END;
);
QUIT;
PROC SQL OUTOBS=10;
SELECT BUSINESS_ID, SUM(WEIGHTEDSCORE)/SUM(W) + RU AS PREDSCORE
FROM USERUSERPREDICTION&I GROUP BY BUSINESS_ID, RU ORDER BY PREDSCORE DESC ;
QUIT;
%END;
%MEND;
%PREDICTSCORE();

Re: COLLABORATIVE-FILTERING SYSTEM

noling — Thu, 29 Aug 2019 13:46:49 GMT

What exactly are you trying to understand? SAS developers don't use the term "Low code commands". All the %DO, %TO, and %END and &&&'s are part of the macro facility, which only generates text, in this case more SAS code.

Looks like the program loops through 2 PROC SQL's I times (15 user input), and loops through several variables J times. It's easier to see if you indent it. Is this helpful?

MACRO PREDICTSCORE(USRCNT=15);
	%DO I = 1 %TO &USRCNT;
		PROC SQL;
			CREATE TABLE USERUSERPREDICTION&I AS
			SELECT BUSINESS_ID, USER_ID, &&RU&I AS RU,
			"&&USER&I" AS USER, CASE USER_ID
			%DO J = 1 %TO &&NCNT_&I;
				WHEN "&&&&NNAME&I._&J." THEN
				&&&&NEIGHBOR&I._&J. * (STARS - &&RN_&I._&J)
			%END;
			END AS WEIGHTEDSCORE,
			CASE USER_ID
			%DO J=1 %TO &&NCNT_&I;
				WHEN "&&&&NNAME&I._&J." THEN &&NEIGHBOR&I._&J.
			%END;
			END AS W
			FROM YELP.USER_PROFILE_V1 WHERE
			USER_ID IN (
			%DO J = 1 %TO &&NCNT_&I;
				"&&&NNAME&I._&J" %IF &J<&USRCNT
			%THEN ,;
			%END;
			);
		QUIT;
		PROC SQL OUTOBS=10;
			SELECT BUSINESS_ID, SUM(WEIGHTEDSCORE)/SUM(W) + RU AS PREDSCORE
			FROM USERUSERPREDICTION&I GROUP BY BUSINESS_ID, RU ORDER BY PREDSCORE DESC ;
		QUIT;
	%END;
%MEND;
%PREDICTSCORE();

Re: COLLABORATIVE-FILTERING SYSTEM

EloarL — Thu, 29 Aug 2019 14:01:50 GMT

I have questions in the following command:

OUTOBS=10

Re: COLLABORATIVE-FILTERING SYSTEM

noling — Thu, 29 Aug 2019 14:04:54 GMT

Oh 🙂

OUTOBS=n

restricts the number of rows (observations) in the output. For example, if you specify OUTOBS=10 and insert values into a table using a query-expression, then the SQL procedure inserts a maximum of 10 rows. Likewise, OUTOBS=10 limits the output to 10 rows.

http://support.sas.com/documentation/cdl/en/proc/61895/HTML/default/viewer.htm#a002473669.htm

Re: COLLABORATIVE-FILTERING SYSTEM

Reeza — Thu, 29 Aug 2019 16:57:17 GMT

Someone is returning the top 10 pred scores by business ID. I think an array would have worked better but whatever works in the end.

Re: COLLABORATIVE-FILTERING SYSTEM

EloarL — Thu, 29 Aug 2019 17:22:25 GMT

Oh yes ... thanks for the personal help!

Re: COLLABORATIVE-FILTERING SYSTEM

ChrisNZ — Fri, 30 Aug 2019 04:17:11 GMT

You might as well learn the proper terms.

This is an option (of the proc sql statement). Not a command.