Querying Shapes of Histories

摘要: We present a shape de nition language, called SDL, for retrieving objects based on shapes contained in the histories associated with these objects. It is small, yet powerful, language that allows rich variety of queries about found historical time sequences. An interesting feature SDL its ability to perform blurry matching. A \blurry" match one where user cares overall but does not care speci c details. Another important e cient implementability. The operators are designed be greedy reduce non-determinism, which turn substantially reduces amount back-tracking implementation. give transformation rules rewriting an expression into more form as well index structure speeding up execution queries. used \mine mined rules" data mining application and capability query behavior over period led discovery new information. Current Address: Politecnico di Milano, Italy. Symbol Description lb ub iv fv slightly increasing transition .05 .19 anyvalue Up highly .20 1.0 down decreasing -.19 -.05 Down -1.0 appears from zero value non-zero 0 nonzero disappears stable nal nearly equal initial -.04 .04 both values Table 1: Illustrative Alphabet