Mining rich network user data to extract valuable information becomes the focus and commercial interest of researchers. Differential privacy protection is a new paradigm based on data distortion, which protects sensitive data while maintains a certain statistical properties by adding random noise, and makes up for the shortcomings that traditional schemas require knowledge background assumptions and cannot analyze quantitatively. In this paper, we propose a DP-MCDBScan schema based on the powerful differential privacy. We perform simulation to evaluate our schema, whose results show that our schema has better efficiency, accuracy, and privacy protection effect than previous schemas.
