Citation Context

...been presented in Goodman et al. (1995). The task of discovering frequent parallel episodes can be stated as a task of discovering all frequent sets, a central phase of discovering association rules (=-=Agrawal et al., 1993-=-). The rule generation methods are essentially the same for association rules and WINEPI. The levelwise main algorithm has also been used successfully in the search of frequent sets (Agrawal et al., 1...

Citation Context

...from a telecommunication network, or different types of user actions, and they have been marked on a time line. Recently, interest in knowledge discovery from sequential data has increased (see e.g., =-=Agrawal and Srikant, 1995-=-; Bettini et al., 1996; Dousson et al., 1993; Hätönen et al., 1996a; Howe, 1995; Jonassen et al., 1995; Laird, 1993; Mannila et al., 1995; Morris et al., 1994; Oates and Cohen, 1996; Wang et al., 1994...

Citation Context

...covery tasks is. There are also some interesting similarities between the286 MANNILA, TOIVONEN AND VERKAMO discovery of frequent episodes and the work done on inductive logic programming (see, e.g., =-=Muggleton, 1992-=-); a noticeable difference is caused by the sequentiality of the underlying data model, and the emphasis on time-limited occurrences. Similarly, the problem of looking for one occurrence of an episode...

Citation Context

... method is not practical for episodes since the deterministic automata could be very large.EPISODES IN EVENT SEQUENCES 287 In stochastics, event sequence data is often called a marked point process (=-=Kalbfleisch and Prentice, 1980-=-). It should be noted that traditional methods for analyzing marked point processes are ill-suited for the cases where the number of event types is large. However, there is a promising combination of ...

Citation Context

...e group of recent events has been studied, e.g., in artificial intelligence, where a similar problem in spirit is the many pattern/many object pattern match problem in production system interpreters (=-=Forgy, 1982-=-). Also, comparable strategies using a sliding window have been used, e.g., to study the locality of reference in virtual memory (Denning, 1968). Our setting differs from these in that our window is a...

Citation Context

...g tools, such as in the TASA system, episodes can provide useful views to the collection of recent alarms. 6. Related work The work most closely related to ours is perhaps (Agrawal and Srikant, 1995; =-=Srikant and Agrawal, 1996-=-). There multiple sequences are searched for patterns that are similar to serial episodes with some extra restrictions and an event taxonomy. Our methods can be extended with a taxonomy by a direct ap...

Citation Context

...extra restrictions and an event taxonomy. Our methods can be extended with a taxonomy by a direct application of the similar extensions to association rules (Han and Fu, 1995; Holsheimer et al.,1995; =-=Srikant and Agrawal, 1995-=-). Also, our methods can be applied on analyzing several sequences; there is actually a variety of choices for the definition of frequency (or support) of an episode in a set of sequences. Patterns ov...

Citation Context

...awal et al., 1993). The rule generation methods are essentially the same for association rules and WINEPI. The levelwise main algorithm has also been used successfully in the search of frequent sets (=-=Agrawal et al., 1996-=-); a generic levelwise algorithm and its analysis has been presented in Mannila and Toivonen (1997). Technical problems related to the recognition of episodes have been researched in several fields. T...

Citation Context

... are similar to serial episodes with some extra restrictions and an event taxonomy. Our methods can be extended with a taxonomy by a direct application of the similar extensions to association rules (=-=Han and Fu, 1995-=-; Holsheimer et al.,1995; Srikant and Agrawal, 1995). Also, our methods can be applied on analyzing several sequences; there is actually a variety of choices for the definition of frequency (or suppor...

Citation Context

...t pattern match problem in production system interpreters (Forgy, 1982). Also, comparable strategies using a sliding window have been used, e.g., to study the locality of reference in virtual memory (=-=Denning, 1968-=-). Our setting differs from these in that our window is a queue with the special property that we know in advance when an event will leave the window; this knowledge is used by WINEPI in the recogniti...

Citation Context

...t the time requirement is roughly linear with respect to the length of the input sequence, as could be expected. Finally, we experimented with protein sequences. We used data in the PROSITE database (=-=Bairoch et al., 1995-=-) of the ExPASy WWW molecular biology server of the Geneva University Hospital and the University of Geneva (ExPASy). PROSITE contains biologically significant DNA and protein patterns that help to id...

Citation Context

...; Laird, 1993; Mannila et al., 1995; Morris et al., 1994; Oates and Cohen, 1996; Wang et al., 1994). One basic problem in analyzing event sequences is to find frequent episodes (Mannila et al., 1995; =-=Mannila and Toivonen, 1996-=-), i.e., collections of events occurring frequently together. For example, in the sequence of figure 1, the episode “E is followed by F” occurs several times, even when the sequence is viewed through ...

Citation Context

...line. Recently, interest in knowledge discovery from sequential data has increased (see e.g., Agrawal and Srikant, 1995; Bettini et al., 1996; Dousson et al., 1993; Hätönen et al., 1996a; Howe, 1995; =-=Jonassen et al., 1995-=-; Laird, 1993; Mannila et al., 1995; Morris et al., 1994; Oates and Cohen, 1996; Wang et al., 1994). One basic problem in analyzing event sequences is to find frequent episodes (Mannila et al., 1995; ...

Citation Context

...and Srikant, 1995; Bettini et al., 1996; Dousson et al., 1993; Hätönen et al., 1996a; Howe, 1995; Jonassen et al., 1995; Laird, 1993; Mannila et al., 1995; Morris et al., 1994; Oates and Cohen, 1996; =-=Wang et al., 1994-=-). One basic problem in analyzing event sequences is to find frequent episodes (Mannila et al., 1995; Mannila and Toivonen, 1996), i.e., collections of events occurring frequently together. For exampl...

Citation Context

...pes of user actions, and they have been marked on a time line. Recently, interest in knowledge discovery from sequential data has increased (see e.g., Agrawal and Srikant, 1995; Bettini et al., 1996; =-=Dousson et al., 1993-=-; Hätönen et al., 1996a; Howe, 1995; Jonassen et al., 1995; Laird, 1993; Mannila et al., 1995; Morris et al., 1994; Oates and Cohen, 1996; Wang et al., 1994). One basic problem in analyzing event sequ...

Citation Context

...nd they have been marked on a time line. Recently, interest in knowledge discovery from sequential data has increased (see e.g., Agrawal and Srikant, 1995; Bettini et al., 1996; Dousson et al., 1993; =-=Hätönen et al., 1996-=-a; Howe, 1995; Jonassen et al., 1995; Laird, 1993; Mannila et al., 1995; Morris et al., 1994; Oates and Cohen, 1996; Wang et al., 1994). One basic problem in analyzing event sequences is to find frequ...

Citation Context

...sed (see e.g., Agrawal and Srikant, 1995; Bettini et al., 1996; Dousson et al., 1993; Hätönen et al., 1996a; Howe, 1995; Jonassen et al., 1995; Laird, 1993; Mannila et al., 1995; Morris et al., 1994; =-=Oates and Cohen, 1996-=-; Wang et al., 1994). One basic problem in analyzing event sequences is to find frequent episodes (Mannila et al., 1995; Mannila and Toivonen, 1996), i.e., collections of events occurring frequently t...

Citation Context

...twork, or different types of user actions, and they have been marked on a time line. Recently, interest in knowledge discovery from sequential data has increased (see e.g., Agrawal and Srikant, 1995; =-=Bettini et al., 1996-=-; Dousson et al., 1993; Hätönen et al., 1996a; Howe, 1995; Jonassen et al., 1995; Laird, 1993; Mannila et al., 1995; Morris et al., 1994; Oates and Cohen, 1996; Wang et al., 1994). One basic problem i...

Citation Context

...ces corresponds here to searching episodes from a sequence of (sender, alarm type) pairs. First order temporal logic has been proposed as a means of both expressing and discovering temporal patterns (=-=Padmanabhan and Tuzhilin, 1996-=-). The formalism is strong and allows expressing more complex patterns than episodes; it is unclear what the complexity of different discovery tasks is. There are also some interesting similarities be...

Citation Context

...e fact that we know where subepisodes of candidates have occurred. The methods for matching sets of episodes against a sequence have some similarities to the algorithms used in string matching (e.g., =-=Grossi and Luccio, 1989-=-). In particular, recognizing serial episodes in a sequence can be seen as locating all occurrences of subsequences, or matches of patterns with variable length do not care symbols, where the length o...

Citation Context

...t in knowledge discovery from sequential data has increased (see e.g., Agrawal and Srikant, 1995; Bettini et al., 1996; Dousson et al., 1993; Hätönen et al., 1996a; Howe, 1995; Jonassen et al., 1995; =-=Laird, 1993-=-; Mannila et al., 1995; Morris et al., 1994; Oates and Cohen, 1996; Wang et al., 1994). One basic problem in analyzing event sequences is to find frequent episodes (Mannila et al., 1995; Mannila and T...

Citation Context

...nd they have been marked on a time line. Recently, interest in knowledge discovery from sequential data has increased (see e.g., Agrawal and Srikant, 1995; Bettini et al., 1996; Dousson et al., 1993; =-=Hätönen et al., 1996-=-a; Howe, 1995; Jonassen et al., 1995; Laird, 1993; Mannila et al., 1995; Morris et al., 1994; Oates and Cohen, 1996; Wang et al., 1994). One basic problem in analyzing event sequences is to find frequ...

Citation Context

...for corrective actions. See, e.g., Jakobson and Weissman (1993) for a description of a representative correlation system; similar approaches have been used successfully also in process control tasks (=-=Milne et al., 1994-=-). While the use of alarm correlation systems is quite popular and methods for specifying the correlations are maturing, acquiring all the knowledge necessary for constructing an alarm correlation sys...

Citation Context

...d on a time line. Recently, interest in knowledge discovery from sequential data has increased (see e.g., Agrawal and Srikant, 1995; Bettini et al., 1996; Dousson et al., 1993; Hätönen et al., 1996a; =-=Howe, 1995-=-; Jonassen et al., 1995; Laird, 1993; Mannila et al., 1995; Morris et al., 1994; Oates and Cohen, 1996; Wang et al., 1994). One basic problem in analyzing event sequences is to find frequent episodes ...

Citation Context

...ences can be within a given edit distance. Recent results on the pattern matching aspects of recognizing episodes can be found in Das et al. (1997). The recent work on sequence data in databases (see =-=Seshadri et al., 1996-=-) provides interesting openings towards the use of database techniques in the processing of queries on sequences. A problem similar to the computation of frequencies occurs also in the area of active ...

Citation Context

...ntial data has increased (see e.g., Agrawal and Srikant, 1995; Bettini et al., 1996; Dousson et al., 1993; Hätönen et al., 1996a; Howe, 1995; Jonassen et al., 1995; Laird, 1993; Mannila et al., 1995; =-=Morris et al., 1994-=-; Oates and Cohen, 1996; Wang et al., 1994). One basic problem in analyzing event sequences is to find frequent episodes (Mannila et al., 1995; Mannila and Toivonen, 1996), i.e., collections of events...