Understand Relevance in Fluid Topics - Fluid Topics - 3.2 - Technical Notes

Understand Relevance in Fluid Topics

Fluid Topics
Fluid Topics Version
Technical Notes
Target Audience

The AFS search solution offers high-level relevance in its services, based on the following components:
  • Lexical analysis algorithms help keep the original forms of the words in addition to the standardized and stemmed forms.
  • Advanced linguistic processes make it possible to extract relevant expressions.
  • The semantic expansion mechanism regulates the relevance level of a word according to its position compared to that of the searched synonym or generic/specific term.
AFS always compares the relative positions of words and the space between them. Searches like "the kings of music" and "the music of kings" are different and processed distinctly to generate more accurate results. Relevance is therefore a result of several steps when AFS parses a user's query:
  1. Analyzing the query
  2. Searching relevant documents in the index
  3. Computing relevance indicators for each result
  4. Sorting results according to criteria selected for relevance