Intelligent search in learning communities

Online learning communities may help increase learning outcome of community members. Intelligent search in those communities may support users to obtain their own needed information. As other social networks there are different types of entities in a learning community, such as learning materials, `similar' persons, learning groups, tags. Accordingly there may be different relationships among them. We argue that those data can be accurately modeled by heterogeneous networks, and explore unified intelligent search that incorporates different types of entities using a spreading activation approach. We propose ISC-Search, which provides users with three types of search possibilities/strategies, in addition to the traditional one, and give them choices to fit their own information needs. These different search strategies leverage social data and content data in different ways to satisfy users' needs. They are (1) Intelligent search, which finds the relevant entities of different types in an online learning community; (2) Social first search, which searches first by social relationships; (3) Content first search, which finds the relevant entities first by content; (4) traditional search, which focuses on the search words themselves. We developed and implemented these search strategies for the online teaching and learning community expertAzubi, and compared these different search strategies in an experiment.