How is relevance determined
Relevance is determined by konfer for a given search term and applied search filters
- Whether a record contains one or more of the individual words in your search query
- How many times each of the words in your query appears in each retrieved resource record, compared to how many times the same words appear across all of the records in Konfer. Words that occur more frequently across the konfer dataset will not be as impactful in terms of relevance
- The length of the field in which a word in your query appears in the resource record. This means that the appearance of a word in a short field, such as a research publication title, leads to the ranking of that resource more highly in the results than the same word appearing in a longer field like the abstract
- Whether words in your query appear adjacent to each other in a single resource field, or if the search words are spread out across the resource record. This fourth factor only applies where multiple words are used in your search query
For experts, relevance will be based on published research for experts within konfer, especially recent research, within your area of interest.
Where the user has filtered a query, then the results are limited to the filtered data. For instance, searches from within an Innovation Hub are limited to the universe of resources for that particular Hub.
Use of the Search Advisor will also limit search results based on the presence of all selected keywords and one or more of each of any topics or concepts selected.
Relevance roundels
Relevance roundels indicate the level of relevance attached to any particular resource by konfer. The roundels are green. Roundels are more complete (as a circle) when relevance is higher. In the example below, the research project is considered by konfer to be of moderate relevance to the search:
A high level of relevance looks like this:
