Why is the automatic product classification not working for my external web content?

Environment

Oracle Knowledge Advanced

Resolution

The automatic product classification process identifies products in external documents and adds the product information to the search index metadata. Knowledge Advanced uses the metadata to classify search results by product and presents aggregated results, through the use of facet widgets, so that users can easily focus on the most relevant answers.  This allows users to use the search filters to narrow the results.

When the product name is found in the following places then the web content will be classified to that product in the product tree.

• matches in the html head tag, title or keyword attributes.
• matches in the body subtitles (h1 through h4 for html documents)
• matches in the document url path
• matches frequently in the body of the page without high references to other products.  This is controlled by a statistical algorithm.

In addition the following applies to classifying the articles. 

• multiple products can be matched up to 50 products.  However when in the body of the document many products can cause no matches because the statistical algorithm determines there is not a strong enough correlation to one or multiple products. 
• The processing will match to the lowest level of the tree.
• Product synonyms do not impact classification, they just impact search results.

If the classification is not appearing as expected you want to check these things -

• The product classification is enabled in the Search Configuration tab.
• The product tree entries are enabled for the interface that you are using to view the search results and filters.
• The specific content has been processed by the content update and then the indexing has been run for any changes that impact the classification.  These can be web document text changes or product hierarchy name changes.