Pattern-based stability analysis of complex data sets
원문보기
IPC분류정보
국가/구분
United States(US) Patent
등록
국제특허분류(IPC7판)
G06F-007/04
G06F-017/30
출원번호
US-0311586
(2011-12-06)
등록번호
US-8874610
(2014-10-28)
발명자
/ 주소
Geroulo, Michael T.
출원인 / 주소
International Business Machines Corporation
대리인 / 주소
Tutunjian & Bitetto, P.C.
인용정보
피인용 횟수 :
1인용 특허 :
3
초록▼
Methods and systems for identifying stability exceptions in a data log are disclosed. In one method, at least one key that is present in the data log is determined. The data log is comprised of at least one data set, at least one of which includes a plurality of iterations indicating states of the c
Methods and systems for identifying stability exceptions in a data log are disclosed. In one method, at least one key that is present in the data log is determined. The data log is comprised of at least one data set, at least one of which includes a plurality of iterations indicating states of the corresponding data set at different points in time. For each data set and for each key, a map is generated. The map indicates, for each iteration of the corresponding data set, whether the corresponding key is present in the corresponding iteration. Moreover, at least one expression pattern rule that models data item stability characteristics over data set iterations is compared to each of the maps to determine whether the corresponding map satisfies the one or more expression pattern rules. Further, at least one unstable data item is identified in the data log based on the comparison.
대표청구항▼
1. A method for identifying stability exceptions in a data log comprising: determining at least one key that is present in the data log, wherein the data log is comprised of at least one data set, at least one of which includes a corresponding plurality of iterations indicating binary transition sta
1. A method for identifying stability exceptions in a data log comprising: determining at least one key that is present in the data log, wherein the data log is comprised of at least one data set, at least one of which includes a corresponding plurality of iterations indicating binary transition states of the corresponding data set at different points in time, and wherein each key of the at least one key denotes a different, respective data item in the data log;for each data set of the at least one data set and for each key of the at least one key, generating a map indicating, for each iteration of the corresponding data set, whether the corresponding key is present in the corresponding iteration;determining adherence, by a processor, of at least one expression pattern rule that models data item stability characteristics over data set iterations to the map generated for each data set and for each key to determine whether the corresponding map satisfies the at least one expression pattern rule; andidentifying at least one key exception as at least one unstable data item, respectively, in the data log based on the determining adherence. 2. The method of claim 1, wherein the method further comprises: outputting a key performance indicator for the data log based on the identifying. 3. The method of claim 1, wherein the method further comprises: outputting a list that includes the at least one unstable data item. 4. The method of claim 1, wherein the at least one expression pattern rule models an unstable data item. 5. The method of claim 4, wherein the at least one expression pattern rule includes a set of consecutive iterations and wherein transitions between iterations in the set of consecutive iterations denote state switches. 6. The method of claim 1, wherein the at least one expression pattern rule models a stable data item. 7. The method of claim 6, wherein the at least one expression pattern rule includes a set of consecutive iterations and wherein each transition between iterations in the set of consecutive iterations denotes a state consistency. 8. A non-transitory computer readable storage medium comprising a computer readable program for identifying stability exceptions in a data log, wherein the computer readable program when executed on a computer causes the computer to perform the steps of: determining at least one key that is present in the data log, wherein the data log is comprised of at least one data set, at least one of which includes a corresponding plurality of iterations indicating binary transition states of the corresponding data set at different points in time, and wherein each key of the at least one key denotes a different, respective data item in the data log;for each data set of the at least one data set and for each key of the at least one key, generating a map indicating, for each iteration of the corresponding data set, whether the corresponding key is present in the corresponding iteration;determining adherence of at least one expression pattern rule that models data item stability characteristics over data set iterations to the map generated for each data set and for each key to determine whether the corresponding map satisfies the at least one expression pattern rule; andidentifying at least one key exception as at least one unstable data item, respectively, in the data log based on the determining adherence. 9. The computer readable storage medium of claim 8, wherein the computer readable program when executed on the computer further causes the computer to perform the step of: outputting a key performance indicator for the data log based on the identifying. 10. The computer readable storage medium of claim 8, wherein the computer readable program when executed on the computer further causes the computer to perform the step of: outputting a list that includes the at least one unstable data item. 11. A method for identifying stability exceptions in a data log comprising: receiving the data log, which is comprised of at least one data set, at least one of which includes a corresponding plurality of iterations indicating binary transition states of the corresponding data set at different points in time, wherein at least one data item in the data log is denoted as a key, wherein each key denotes a different, respective data item in the data log;for each of the keys, determining a data coverage of the corresponding key for each iteration of the data log, wherein the data coverage denotes a degree to which the corresponding key constitutes the data log in a given iteration;for each of the keys, generating a map indicating whether the data coverage of the corresponding key has changed between iterations of the data log; anddetermining adherence, by a processor, of at least one expression pattern rule that models data item stability characteristics over data log iterations to the map generated for each data set and for each key to identify at least one key exception as at least one unstable data item, respectively, in the data log. 12. The method of claim 11, wherein the method further comprises: outputting, based on the determining adherence, at least one of a key performance indicator for the data log or a list that includes the at least one unstable data item. 13. The method of claim 11, wherein each of the maps indicates, for each transition between iterations of the data log, whether the data coverage of the corresponding key has increased, decreased or remained unchanged for the corresponding transition. 14. The method of claim 11, wherein the at least one expression pattern rule models an unstable data item. 15. The method of claim 14, wherein the at least one expression pattern rule includes elements that denote transitions between iterations and wherein the elements denote at least two state switches between an increasing data coverage and a decreasing data coverage of the unstable data item modeled by the at least one expression pattern rule. 16. The method of claim 11, wherein the at least one expression pattern rule models a stable data item. 17. The method of claim 16, wherein the at least one expression pattern rule includes a set of elements that denotes consecutive transitions between iterations and wherein the set of elements denotes a state consistency between the consecutive iterations. 18. A system for identifying stability exceptions in a data log comprising: a parser configured to receive the data log, which is comprised of at least one data set, at least one of which includes a corresponding plurality of iterations indicating binary transition states of the corresponding data set at different points in time, wherein at least one data item in the data log is denoted as a key, wherein each key denotes a different, respective data item in the data log;a data encoder configured to determine, for each of the keys, a data coverage of the corresponding key for each iteration of the data log, wherein the data coverage denotes a degree to which the corresponding key constitutes the data log in a given iteration, and configured to generate, for each of the keys, a map indicating whether the data coverage of the corresponding key has changed between iterations of the data log; andprocessor, configured to determine adherence of at least one expression pattern rule that models data item stability characteristics over data log iterations to the map generated for each data set and for each key to identify at least one key exception as at least one unstable data item, respectively, in the data log. 19. The system of claim 18, further comprising: a computation module configured to calculate and output a key performance indicator based on a determination of adherence between the at least one expression pattern rule and the map generated for each data set and for each key. 20. The system of claim 18, wherein each of the maps indicates, for each transition between iterations of the data log, whether the data coverage of the corresponding key has increased, decreased or remained unchanged for the corresponding transition. 21. The system of claim 18, wherein the at least one expression pattern rule models an unstable data item, wherein the at least one expression pattern rule includes elements that denote transitions between iterations and wherein the elements denote at least two state switches between an increasing data coverage and a decreasing data coverage of the unstable data item modeled by the at least one expression pattern rule. 22. The system of claim 18, wherein the at least one expression pattern rule models a stable data item, wherein the at least one expression pattern rule includes a set of elements that denotes consecutive transitions between iterations and wherein the set of elements denotes a state consistency between the consecutive iterations. 23. The system of claim 18, further comprising: a key encoder configured to generate, for each data set of the at least one data set and for each of the keys, another map indicating, for each iteration of the corresponding data set, whether the corresponding key is present in the corresponding iteration; anda key matcher configured to determining adherence of at least one other expression pattern rule that models data item stability qualities over data set iterations to the other map generated for each data set of the at least one data set and for each of the keys to determine whether the corresponding other map satisfies the at least one other expression pattern rule. 24. The system of claim 23, wherein the at least one other expression pattern rule models an unstable data item, wherein the at least one other expression pattern rule includes a set of consecutive iterations and wherein transitions between iterations in the set of consecutive iterations denote state switches. 25. The system of claim 23, wherein the at least one other expression pattern rule models a stable data item, wherein the at least one expression pattern rule includes a set of consecutive iterations and wherein each transition between iterations in the set of consecutive iterations denotes a state consistency.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (3)
Rorvig Mark E. (Houston TX), General method of pattern classification using the two domain theory.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.