Analysis system and method for intelligent customer service based on common sequence pattern
The present invention provides a method and system for information analysis. The method extracts a plurality of metadata from a source information set so as to generate a metadata set, the metadata comprising error log information that forms an error sequence in the metadata set; and generates a common error sequence set for a plurality of the metadata sets. By means of the method, it is possible to easily obtain a global error sequence pattern, and easily compare a new error with a previous error sequence pattern so as to prevent, diagnose and recover the new error.
1. A computer-executable method for information analysis using an analysis system including one or more processors, comprising:
extracting, using the one or more processors, metadata from a source information set creating a metadata set, the metadata comprising error log information that forms an error sequence in the metadata set;
generating a common error sequence set for the metadata set by comparing the metadata set to a previously generated metadata set;
filtering invalid common error sequences, wherein the analysis system is enabled to place invalid common error sequences in a pattern blacklist for future filtering;
periodically analyzing the pattern blacklist to determine whether valid common error sequences were placed in the pattern blacklist; and
providing and storing the common error sequence set in a repository to enable diagnosis of future errors.
2. The computer-executable method according to claim 1 , further comprising:
wherein the metadata comprises time information; and
normalizing the time information in the metadata set.
3. The computer-executable method according to claim 1 , further comprising:
de-duplicating error log information in the metadata set.
4. The computer-executable method according to claim 3 , wherein the de-duplicating error log information comprises removing messages of one or more types of:
totally the same error logs, error logs that are totally the same except the non-critical parts, error logs with the same critical parts, and user-confirmed duplicated error logs.
5. The computer-executable method according to claim 3 , wherein the de-duplicating error log information comprises implementing by machine learning and/or users.
6. The computer-executable method according to claim 1 , wherein the extracted metadata is metadata in a defined source.
7. The computer-executable method according to claim 1 , wherein the generating a common error sequence set for a plurality of the metadata sets comprises:
generating a common error sequence set for every two metadata sets of the plurality of the metadata sets; and
merging the common error sequence sets generated for every two metadata sets into a common error sequence set of the plurality of the metadata sets.
8. The computer-executable method according to claim 7 , wherein the generating a common error sequence set for every two metadata sets of the plurality of the metadata sets is implemented in parallel.
9. The computer-executable method according to claim 7 , wherein the merging the common error sequence sets generated for every two metadata sets into a common error sequence set of the plurality of the metadata sets further comprises:
making statistics regarding how many times each common error sequence occurs.
10. The computer-executable method according to claim 1 , wherein the error sequence identifies relations between the error log information.
11. The computer-executable method according to claim 1 , wherein the source information comprises information related to one or more of: application logs, system events, system logs, process dumps, thread dumps, user records, and application records; and/or the error log information comprises information related to one of more of: runtime error, warn, and DEBUG.
12. The computer-executable method according to claim 1 , wherein the metadata further comprises log level and/or sources of the error log information.
13. The computer-executable method according to claim 1 , wherein the common error sequence set comprises an error sequence set formed by the same error log information in the plurality of the metadata sets.
14. The computer-executable method according to claim 13 , wherein a length of the error sequence formed by the same error log information is longer than a defined threshold.
15. The computer-executable method according to claim 13 , wherein a judgment criterion for being the same is defined by users.
16. A system, comprising:
an analysis system including one or more processors; and
computer-executable program logic encoded in memory of one or more computers enabled to analyze information using the analysis system, wherein the computer-executable program logic is configured for the execution of:
extracting, using the one or more processors, metadata from a source information set creating a metadata set, the metadata comprising error log information that forms an error sequence in the metadata set;
generating a common error sequence set for the metadata set by comparing the metadata set to a previously generated metadata set;
filtering invalid common error sequences, wherein the analysis system is enabled to place invalid common error sequences in a pattern blacklist for future filtering;
periodically analyzing the pattern blacklist to determine whether valid common error sequences were placed in the pattern blacklist; and
providing and storing the common error sequence set in a repository to enable diagnosis of future errors.
17. The system according to claim 16 , wherein the metadata further comprises time information, and the wherein the computer-executable program logic is further configured for the execution of normalizing the time information in the metadata set.
18. The system according to claim 16 , wherein the computer-executable program logic is further configured for the execution of:
de-duplicating error log information in the metadata set.
19. The system according to claim 18 , wherein the de-duplicating error log information comprises removing messages of one or more types of:
totally the same error logs, error logs that are totally the same except the non-critical parts, error logs with the same critical parts, and user-confirmed duplicated error logs.
20. The system according to claim 18 , wherein the de-duplicating error log information comprises implementing by machine learning and/or users.