With the aid of information communications technology, telemedicine can break the limitation of physical space, which has a promising future. Nowadays, there are numerous telemedicine systems, but most of them lack the active interaction with users. Under this background, based on data mining, we construct push service for telemedicine system, which includes disease differentiation, doctor recommendation and diagnosis prediction, and design k-nearest neighbor classification, nearest neighbor recommendation and FP-growth as main algorithms of them respectively. Moreover, we carry out an empirical research by using data from a public Chinese telemedicine system.
