Most commonly-used techniques in smart environments such as ADL recognition are designed and tested for a specific space and a specific person; therefore learning in each environmental situation is treated as a separate context. In this paper, we try to develop a method for recognizing and transferring learned knowledge of activities between different residents. Our method is able to map activities despite intra-subject variability and inter-subject variability, by using a discontinuous mining method and a similarity measurement method. At the end, we will provide the results of our experiments on real data obtained from a smart apartment.