Content analysis - given terms relevance score

Question

We are thinking to use the existing Content analysis service API to return the given terms relevance score in the article.

Found some very good content analysis service such as:

1. Yahoo Content analysis service
2. OpenCalais ...

However, we are looking for a way to return the given terms relevance score

for example,

we have a vocabulary - topics contains "cost", "government"

and a page http://www.test.dev/article/1.html

then we want to use the content analysis service to check the relevance for the topics in the article

so we can decide to auto tag the article from our own vocabulary.

thanks for the help.

Answer 1

The Yahoo Content Analysis service you mention would be useful for identifying related links and other metadata. If you were specifically looking for vocabulary terms, you could try the Yahoo Term Extraction service. It accepts the source text and query terms.

Here's the demo query as shown in the YQL Console:

select * from search.termextract where context="Italian sculptors and painters of the renaissance favored the Virgin Mary for inspiration" and query="madonna"

And sample results:

"results": {
 "Result": [
  "Italian sculptors",
  "the Virgin Mary",
  "painters",
  "the renaissance"
 ]
}

One large caveat: Yahoo has said the Term Extraction service will be shut down at some point. Instead they are recommending new development use the Content Analysis service. (Source)

There are also YQL tables for OpenCalais if you want to try that service results in the same context.