Research on User Identification across Multiple Social Networks Based on Preference

In the view of the lack of user’s property information and low utilization rate of the user’s dynamic behavior data in the networks, a new identification method that incorporate user’s dynamic preferences (DPUI) is proposed. First, we analyze nickname’s naming features on massive real data sets and constructed a nickname similarity calculation approach. Then, we use time sliding window and LDA model to mine user’s long-term dynamic preferences from their behavior data. Finally, integrated user’s preference to identify users. Experimental results show that the proposed algorithm can effectively identify user’s identity with high precision and recall rate.