Live video recommendation by an online system

A non-transitory computer readable medium storing executable computer program instructions, the computer program instructions comprising instructions that when executed cause a computer processor to:

• receive a plurality of live videos for streaming to client devices associated with users of an online system;
• receive a request for content from a target user during the streaming of the live videos;
• in response to the request for content, dynamically identify a topic for each of the plurality of live videos during the streaming of the live video;
• select a group of live videos as candidates for presentation to the target user, the selection based on identified topics of the candidates matching one or more interests of the target user;
• for each candidate, determine a ranking score indicating a probability that the target user would positively interact with the candidate if the candidate is presented to the target user in the newsfeed;
• rank the candidates based on the ranking scores of the candidates;
• provide, for display to the target user, one or more of the candidates that are above a threshold level in the ranking.

A computer system comprising:

• a non-transitory computer-readable storage medium storing executable computer program instructions, the computer program instructions comprising instructions that when executed cause a computer processor to perform steps, comprising:
• receiving a plurality of live videos for streaming to client devices associated with users of an online system;
• receiving a request for content from a target user during the streaming of the live videos;
• in response to the request for content, dynamically identifying a topic for each of the plurality of live videos during the streaming of the live video;
• selecting a group of live videos as candidates for presentation to the target user, the selection based on identified topics of the candidates matching one or more interests of the target user;
• for each candidate, determining a ranking score indicating a probability that the target user would positively interact with the candidate if the candidate is presented to the target user in the newsfeed;
• ranking the candidates based on the ranking scores of the candidates;
• providing, for display to the target user, one or more of the candidates that are above a threshold level in the ranking.