+ Start a Discussion
Sammi SibtainSammi Sibtain 

Salesforce Knowledge: articleList Keyword relevance

Hi,

As per the documentation, if you give a keyword in VF article list tag, the article that will be displayed on page will be sorted based on keyword relevance. I don't understand what it is meant with the "keyword relevance". What does that mean concrete? What are the criterias to be see as "related article" ?

Thanks
Best Answer chosen by Sammi Sibtain
KevinPKevinP
Sammi,

Think of this keyword relevance as a bit of fuzzy logic, with some concrete bounds. The general idea, is that each article will have more than one tag. Lets say in our hypothetical world here, that each article has 10 tags, each of which is a color. Looking at an article about red-blue-green widgets you might expect it to be tagged with red, blue and green. 

A searcher might search for red and blue widgets. As a result, related articles would likely show articles about "blue widgets" and "red widgets". 
However, it would also show articles that matched more than one keyword, assuming they were more relevant, becuse more more keywords matched the search.  Thus, in our fictional search for red and blue widgets, we most likely see the article for red widets, blue widgets but also red-green-blue widgets! Additionally, it may place the red-green-blue higher up the list, because more than one keyword matches. 

Unfortunately, at the end of the day, the algorithm for keyword relevance is likely a deeply held secret. The above is just a general synopsis of how those algorithms work. 

All Answers

KevinPKevinP
Sammi,

Think of this keyword relevance as a bit of fuzzy logic, with some concrete bounds. The general idea, is that each article will have more than one tag. Lets say in our hypothetical world here, that each article has 10 tags, each of which is a color. Looking at an article about red-blue-green widgets you might expect it to be tagged with red, blue and green. 

A searcher might search for red and blue widgets. As a result, related articles would likely show articles about "blue widgets" and "red widgets". 
However, it would also show articles that matched more than one keyword, assuming they were more relevant, becuse more more keywords matched the search.  Thus, in our fictional search for red and blue widgets, we most likely see the article for red widets, blue widgets but also red-green-blue widgets! Additionally, it may place the red-green-blue higher up the list, because more than one keyword matches. 

Unfortunately, at the end of the day, the algorithm for keyword relevance is likely a deeply held secret. The above is just a general synopsis of how those algorithms work. 
This was selected as the best answer
Sammi SibtainSammi Sibtain

Kevin,

Thanks for your detailed response. What I understood is prefered search term should be effectively used by the users in order to decide the keyword relevance and to decide the sorting crieteria in PKB search.

Just to make sure we are on the same page, we use prefered search term to tag an article, correct? Below are the links.

http://releasenotes.docs.salesforce.com/en-us/winter15/release-notes/rn_promoted_search_terms.htm

https://developer.salesforce.com/docs/atlas.en-us.api.meta/api/sforce_api_objects_searchpromotionrule.htm

Thanks again,

Sammi

 

KevinPKevinP
Thats correct!