There are four keyword recommendation scenarios in the QQ browser. The first one is personalized word recommendation which recommends keywords that users are interested in based on their interest preferences. The second one is automated complete word recommendation. After users enter a particular keyword in the search box, the system will automatically complete related search terms. At this time, users can click on the recommended search terms. The third is based on the articles the user has read, and then recommend article-related search terms. The fourth one is to display other search terms related to the currently entered search term on the search result page according to the search term entered by the user.

How are the search terms generated?

1. Keywords entered by the user, also called active search words. Users normally do not enter complete sentences, thus resulting in inaccurate recommendations;
2. Replace the core words. 
3. Based on the internal knowledge.
4. Based on popular search, The thesaurus system will determine the quality of words, access the security system, and the review mechanism.

How to evaluate the search recommendation?

Relevancy. First of all, look at the relevance indicators to see whether the search results meet the needs of users. This can be seen by the click rate of users on search terms;
Completeness of information. The search results should be complete.

Timeliness. Some categories, such as stock prices, news, and weather, must present the latest search results to users.