Topic popularity prediction method for incremental group level of online social network

The invention provides a topic popularity prediction method for an incremental group level of an online social network. The prediction method comprises the following steps: collecting data; dividing users into different user groups according to the collected user network structure data and user forwarding behavior data, calculating the similarity among topics according to the collected historical topic propagation data, and selecting Top-K similar topics for the predicted target topic; according to the Top-K similar topics, the user group and popularity values at different times, constructing a group level popularity tensor; carrying out incremental prediction on the popularity tensor of the group level by adopting incremental CP decomposition; and restarting CP decomposition to reduce accumulative errors. Compared with the prior art, the topic popularity prediction method for the incremental group level of the online social network is higher in prediction efficiency and prediction precision.