DBLP is popular platform to manage publications and researchers. When analysing relationship between researchers, we often use the public data from DBLP to model a bibliographic network where each node is a researcher, and two researcher are connected by an edge if they have a common publication. Community detection and community search are important problems in data mining. In this project, we aim to use cohesive subgraph models (e.g., k-core, k-truss, k-plex...) to mine real research communities from DBLP. To do this project, you need to consider how to organise and preprocess the graph data on the backend for efficient query processing. An example can be found in the Figure 11 of the following paper:
Data mining | Big data | Community detection | Graph analysis | Data structure and algorithms
The research will be conducted in the DKR group (https://unswdb.github.io/). Further HDR (MPhil/PhD) positions and scholarships are available for excellent project students.
A user interface (e.g., a website) to display the result community queried by the user. The output is supported by carefully designed algorithms and data processing techniques.