- Understanding Custom Context
- Leveraging Custom Contexts
- Feature Selection
- Event Recommendations Deployment Overview
- Deploying Dynamic Navigation Experience
- Helping Train ART Models by Linking Queries to Results
- Adding Coveo Machine Learning Blacklist Words
- About the PermanentId Field
Adding Coveo Machine Learning Blacklist Words
Depending on the nature of the user queries going through your Coveo-powered search interfaces, Coveo Machine Learning models can ignore search requests containing blacklisted words. Those queries are not taken into account during the model learning process, meaning that models will never suggest them. However, in specific scenarios such as support services for youth, it may not be necessary to blacklist some words. In such cases, you can contact your Coveo Customer Success Manager (CSM) to review Coveo’s blacklist and potentially remove some entries. You can also contact Coveo Support to send your own text file of blacklist words.
You might also consider adding stop word rules to exclude bad keywords from queries to not impact the Coveo ML model learning process. However, if end users perform queries containing only banned words, the model learning process could be affected depending on the returned search results (if any) (see Stop Word Special Cases).
To prevent users from getting search results when they input queries that contain only stop words, you can add thesaurus rules in which a given banned keyword is replaced with random characters (see Adding and Managing Query Pipeline Thesaurus Rules). This should bring the user to a “No Results” page where the
QuerySummary component displays the desired information (see Displaying Content Based on Whether There Are Query Results).
As this impacts metrics such as Search Events Without Results and Search Event Click-Through, ensure that the thesaurus rules are not excluding desired results.
You set a stop words rule to ban the word
knight because it leads to unwanted search results.
Knowing that the word
knight queried by itself still leads to the unwanted search results, due to a stop word rule behavior, you set the following thesaurus rule:
knight is now replaced with
asdfgh which leads users to a “No Results” page when they search for it.
In the no results page, you set the following information to be displayed to users who queried