Interactive sequential pattern mining
Interactive sequential pattern mining is disclosed. One example is a system including a sequence miner, and an interaction processor. A sequence database is received, the sequence database including a plurality of input sequences, where each sequence of the plurality of input sequences is an ordered list of events, and each event in the list of events includes at least one item. The sequence miner mines the sequence database for a plurality of candidate sequence patterns, the mining based on an interaction with a user. The interaction processor processes the interaction with the user, the interaction based on domain relevance of the plurality of candidate sequence patterns to the user.
1. A system comprising:
a processor;
a sequence database including a plurality of input sequences, wherein each sequence of the plurality of input sequences is an ordered list of events, and each event in the ordered list of events includes at least one item;
a non-transitory computer readable storage medium storing instructions executable by the processor to:
mine the sequence database for a plurality of candidate sequence patterns, the mining based on an interaction with a user, wherein to mine the sequence database, the processor is to:
determine if the plurality of candidate sequence patterns is a null set;
generate and rank a first plurality of sequence patterns by descending support count when the plurality of candidate sequence patterns is a null set;
determine if a constraint set is an empty set when the plurality of candidate sequence patterns is not a null set;
generate and rank a second plurality of sequence patterns by descending support count when the constraint set is an empty set; and
generate and rank a third plurality of sequence patterns that satisfy each constraint in the constraint set when the constraint set is not an empty set;
provide one of the first, second and third ranked plurality of sequence patterns for the interaction with the user; and
process, via the processor, the interaction with the user, the interaction based on a domain relevance of the one of the first, second and third ranked plurality of sequence patterns provided to the user.
2. The system of claim 1 , wherein the instructions stored on the non transitory computer readable storage medium are executable by the processor to:
provide, to the user, an interactive visual representation of the provided one of the first, second and third ranked plurality of sequence patterns.
3. The system of claim 1 , wherein the instructions stored on the non transitory computer readable storage medium are executable by the processor to:
receive, from the user, an indication that the user is satisfied with at least one sequence pattern of the provided one of the first, second and third ranked plurality of sequence patterns.
4. The system of claim 1 , wherein the instructions stored on the non-transitory computer readable storage medium are executable by the processor to:
receive, from the user, an indication that the user is not satisfied with any sequence pattern of the provided one of the first, second and third ranked plurality of sequence patterns.
5. The system of claim 4 , wherein the instructions stored on the non-transitory computer readable storage medium are executable by the processor to:
receive, from the user, a constraint to guide the mining of the sequence database.
6. The system of claim 5 , wherein the constraint is at least one of a logical constraint and a temporal constraint.
7. The system of claim 4 , wherein the instructions stored on the non-transitory computer readable storage medium are executable by the processor to:
receive, from the user, at least one modified sequence pattern from the provided one of the first, second and third ranked plurality of sequence patterns; and mine the sequence database based on the at least one modified sequence pattern.
8. The system of claim rare input data is received via the processor, and further including a data converter to generate the sequence database from the raw input data.
9. The system of claim 1 , wherein a given sequence of the plurality of input sequences is associated with at least one of:
a sequence identifier identifying the given sequence,
an event identifier identifying a given event in the ordered list of events associated with the given sequence,
an event time identifier identifying a time when the given event occurred, and
an item identifier associated with the at least one item in the given event.
10. A method to find sequential patterns in a sequence database, the method comprising:
mining, via a processor, a sequence database for a plurality of candidate sequence patterns, the sequence database including a plurality of input sequences, wherein each sequence of the plurality of input sequences is an ordered list of events, and each event in the ordered list of events includes at least one item,
wherein the mining of the sequence database further includes:
determining if the plurality of candidate sequence patterns is a null set; and
generating and ranking a first plurality of sequence patterns by descending support count when the plurality of candidate sequence patterns is a null set;
determining if a constraint set is an empty set when the plurality of candidate sequence patterns is not a null set:
generating and ranking a second plurality of sequence patterns by descending support count when the constraint set is an empty set; and
generating and ranking a third plurality of sequence patterns that satisfy each constraint in the constraint set when the constraint set is not an empty set;
providing to a user, via the processor, an interactive visual representation of one of the first, second and third ranked plurality of sequence patterns; and
determining a domain relevance of the one of the first, second and third ranked plurality of sequence patterns provided to the user, the determination based on interacting with the user via the interactive visual representation.
11. The method of claim 10 , wherein the determining the domain relevance includes receiving, from the user, an indication that the user is satisfied with at least one sequence pattern of the provided one of the first, second and third ranked plurality of sequence patterns.
12. The method of claim 10 , wherein the determining the domain relevance includes receiving, from the user, an indication that the user is not satisfied with any sequence pattern of the provided one of the first, second and third ranked plurality of candidate patterns.
13. A non-transitory computer readable medium comprising machine readable instructions executable by a processor to:
receive a sequence database, the sequence database including a plurality of input sequences, wherein each sequence of the plurality of input sequences is an ordered list of events, and each event in the ordered list of events includes at least one item;
mine the sequence database for a plurality of candidate sequence patterns based on an interaction with a user, wherein to mine the sequence database, the processor is to:
determine if the plurality of candidate sequence patterns is a null set; and
generate and rank a first plurality of sequence patterns by descending support count when the plurality of candidate sequence patterns is a null set;
determine if a constraint set is an empty set when the plurality of candidate sequence patterns is not a null set;
generate and rank a second plurality of sequence patterns by descending support count when the constraint set is an empty set; and
generate and rank a third plurality of sequence patterns that satisfy each
constraint in the constraint set when the constraint set is not an empty set; and
process the interaction with the user, the interaction based on an interactive visual representation and a domain relevance of one of the first, second and third ranked plurality of candidate-sequence patterns.
14. The non-transitory computer readable medium of claim 13 , wherein:
to process the interaction with the user, the processor is to receive a constraint from the user; and
to mine, the processor is to mine the sequence database based on the constraint.
15. The method of claim 12 , further including:
receiving, from the user, at least one modified sequence pattern; and
mining the sequence database based on the at least one modified sequence pattern.
16. The method of claim 10 , wherein a given sequence of the plurality of input sequences is associated with at least one of:
a sequence identifier identifying the given sequence,
an event identifier identifying a given event in the ordered list of events associated with the given sequence,
an event time identifier identifying a time when the given event occurred, and
an item identifier associated with the at least one item in the given event.
17. The non-transitory computer readable medium of claim 13 , wherein the processor is to further execute the machine-readable instructions to:
receive, from the user, an indication that the user is satisfied with at least one sequence pattern of the one of the first, second and third ranked plurality of sequence patterns.
18. The non-transitory computer readable medium of claim 13 , wherein the processor is to further execute the machine-readable instructions to:
receive, from the user, an indication that the user is not satisfied with any sequence pattern of the one of the first, second and third ranked plurality of sequence patterns.
19. The non-transitory computer readable medium of claim 18 , wherein the processor is to further execute the machine-readable instructions to:
receive, from the user, at least one modified sequence pattern; and
mine the sequence database based on the at least one modified sequence pattern.