Telegram group recommendation based on users' migration

Today, social networks and messengers have attracted the attention of many different businesses. Every day, a lot of information is produced in these environments. Analyzing this information is very useful for connecting different businesses. This information is very valuable for marketers to find the target community. Telegram is a messenger based on cloud computing. This messenger is used as a social network in some countries, including Iran. Telegram, while used as a social network, does not offer all the capabilities of a social network. The capabilities provided in this messenger include creating a channel, group, and bot. The shortfall in most messengers, such as Telegram, is the limited search service of groups and a community of users. In this paper, we have recommended groups according to the users ' interests, using the graph of users' membership and analyzing their membership records. The proposed method, considering the users' status, models their records in each group. We obtained users’ migration by analyzing their records in each group. Users' migration is analyzed based on the maximum number of users leaving each group and entering another group. In this study, information about 70 million users and 700,000 Telegram supergroups have been used. The evaluation of the proposed model has been done on 30 high-quality groups in Telegram. Selected groups had between 5,000 and 15,000 members. The proposed method showed an error reduction of 0.0237 in RMSE compared to a base method.