Invitation behavior prediction method and device

The invention relates to an invitation behavior prediction method and device and belongs to the field of Internet technology. The method comprises the steps that group behavior feature information ofa first user is acquired; group relation feature information of a second user is acquired; group architecture feature information of a group, the group behavior feature information of the first user and the group relation feature information of the second user are input into an invitation prediction model to obtain a target member user and candidate invitation users of the target member user; andinvitation prediction information is sent to the target member user, wherein the invitation prediction information is used for prompting the target member user to add the candidate invitation users ofthe target member user into the group. According to the method, invitation behavior prediction is performed through the invitation prediction model trained based on features capable of representing the relation among member users in the group, invitation behaviors of the member users and the like, so that the intelligence of a social application is optimal, and the utilization rate of an interactive function provided by the group in the social application is maximized.