Tip of the Day: How to Use Boosting in a Search Query

Searching, Tip of the Week, Tips and Tricks
HeinOnline Blogger

Have you ever wanted to place more relevance on a single term or phrase within your search query? If so, you can use boosting to influence the relevancy ranking of your search results.

Check out the Tip of the Day via our YouTube video, or continue below the video to read a full-text version of this tip.

For example, if you are searching for three terms but want to place more importance and relevance on one of the terms, you can boost the term using the carot symbol ^ (Shift + 6 on your keyboard).

Many often ask, what is the range for boosting, do the numbers go between 1 and 10 or 1 and 100? And how does the boosting algorithm work? For example, if you boost a term by 5, does that mean the search term will appear five times more often?

So, let’s explain the logic behind boosting.Boosting is effectively applying a multiplier to the relevancy ranking given to the boosted term in a specific query. When a search is performed each result is given a score for how well it fulfills the query as a whole. The results are then ranked and displayed based on this scoring. Adding boosting to a term causes the results with that term to be scored higher, and thus displayed higher in the relevancy ranking. It does not change the results you receive, it just changes the order in which they appear when they are sorted using relevance.

Furthermore, there is not a specific allowable range for boosting, as its usage can vary based on the size and scope of different queries. However, numbers in the range of 1-5 typically produce the more recognizable results.

Let’s look at an example, to see how this works.

We want to find documents that discuss ObamaCare, the “individual mandates” that are part of that plan as well as the discussion surrounding medicare.

Obamacare AND “individual mandate*” AND medicare

This search produces 157 results. So we can look at the view matching text pages of our first few results to see our terms. Notice in the first few results that it contains all of our search terms from our search query.  Pay particular attention to how often medicare shows up in the text of the first few results.  Now, let’s manipulate the relevancy ranking and pull to the top of our results list articles that discuss in more detail obamacare as it pertains to medicare. So we are going to boost our term medicare in our search.

Obamacare AND “individual mandate*” AND medicare^5

Now you’ll notice that we still have 157 results, but the results that are now at the top of our list are different. And using VMTP, we see that these articles more heavily discuss medicare than in our previous results.

So we can use boosting to influence the relevancy ranking of our search results list and bring articles and documents to the top of our list that discuss a term more heavily than another in our query.

 

Leave a Reply

Your email address will not be published. Required fields are marked *