Method and system for log based issue prediction using SVM+RNN artificial intelligence model on customer-premises equipment
A method, a set-top box, and a non-transitory computer readable medium for log based issue prediction. The method includes receiving, on a processing server, system log files from a customer-premises equipment, the system log files containing events that are logged by an operating system of the customer-premises equipment; parsing, by the processing server, the events of the system log files to processes and mapping the processes to one or more components of the customer-premises equipment; extracting, by the processing server, features from the mapped processes of the one or more components of the customer-premises equipment; classifying, by the processing server, the extracted features with a first machine learning algorithm; and predicting, by the processing server, anomalies in one or more components of the customer-premises equipment with a second machine learning algorithm using the classified features from the first machine learning algorithm.
1. A method for log based issue prediction, the method comprising:
receiving, on a processing server, system log files from a customer-premises equipment, the system log files containing events that are logged by an operating system of the customer-premises equipment;
parsing, by the processing server, the events of the system log files to processes and mapping the processes to one or more components of the customer-premises equipment;
extracting, by the processing server, features from the mapped processes of the one or more components of the customer-premises equipment;
classifying, by the processing server, the extracted features with a first machine learning algorithm;
predicting, by the processing server, anomalies in one or more components of the customer-premises equipment with a second machine learning algorithm using the classified features from the first machine learning algorithm; and
executing, by the processing server, a fix of the anomalies in the one or more components of the customer-premises equipment.
2. The method according to claim 1 , wherein the first machine learning algorithm is a support vector learning algorithm, and the second machine learning algorithm is a recurrent neural network algorithm.
3. The method according to claim 1 , further comprising:
training the first machine learning algorithm to a first predetermined accuracy; and
training the second machine learning algorithm to a second predetermined accuracy.
4. The method according to claim 3 , further comprising:
separating, by the processing server, each of the features from the mapped processes into an error log entry or a normal log entry for the training of the first machine learning algorithm.
5. The method according to claim 1 , wherein the predicted anomalies in the features of the one or more components of the customer-premises equipment using the second machine learning algorithm comprises a defect in code of one or more of the components of the customer-premises equipment.
6. The method according to claim 1 , wherein the customer-premises equipment is a set-top box.
7. The method according to claim 1 , further comprising:
performing, by the processing server, the parsing of the events of the system log files to the processes and mapping the processes to one or more components of the customer-premises equipment with an algorithm for log parsing.
8. The method according to claim 1 , wherein the first machine learning algorithm converts the system log files into a vector matrix using complete log messages rather than keywords.
9. The method according to claim 1 , wherein the processing server is a part of the customer-premises equipment.
10. The method according to claim 1 , wherein the processing server is a part of a cable provider server.
11. A set-top box comprising:
a processor configured to:
receive system log files containing events that are logged by an operating system of the set-top box;
parse the events of the system log files to processes and mapping the processes to one or more components of the set-top box;
extract features from the mapped processes of the one or more components of the set-top box;
classify the extracted features with a first machine learning algorithm;
predict anomalies in one or more components of the set-top box with a second machine learning algorithm using the classified features from the first machine learning algorithm; and
execute a fix of the anomalies in the one or more components of the set-top box.
12. The set-top box according to claim 11 , wherein the first machine learning algorithm is a support vector learning algorithm, and the second machine learning algorithm is a recurrent neural network algorithm.
13. The set-top box according to claim 11 , wherein the first machine learning algorithm is trained to a first predetermined accuracy, and the second machine learning algorithm is trained to a second predetermined accuracy.
14. The set-top box according to claim 13 , wherein the predicted anomalies in the features of the one or more components of the set-top box comprises a defect in code of one or more of the components of the set-top box.
15. The set-top box according to claim 11 , wherein the processor is further configured to:
perform the parsing of the events of the system log files to the processes and mapping the processes to one or more components of the set-top box with an algorithm for log parsing.
16. A non-transitory computer readable medium having instructions operable to cause one or more processors to perform operations comprising:
receiving, on a processing server, system log files from a customer-premises equipment, the system log files containing events that are logged by an operating system of the customer-premises equipment;
parsing, by the processing server, the events of the system log files to processes and mapping the processes to one or more components of the customer-premises equipment;
extracting, by the processing server, features from the mapped processes of the one or more components of the customer-premises equipment;
classifying, by the processing server, the extracted features with a first machine learning algorithm;
predicting, by the processing server, anomalies in one or more components of the customer-premises equipment with a second machine learning algorithm using the classified features from the first machine learning algorithm; and
executing, by the processing server, a fix of the anomalies in the one or more components of the customer-premises equipment.
17. The non-transitory computer readable medium according to claim 16 , wherein the first machine learning algorithm is a support vector learning algorithm, and the second machine learning algorithm is a recurrent neural network algorithm.
18. The non-transitory computer readable medium according to claim 16 , further comprising:
training the first machine learning algorithm to a first predetermined accuracy; and
training the second machine learning algorithm to a second predetermined accuracy.
19. The non-transitory computer readable medium according to claim 18 , further comprising:
separating, by the processing server, each of the features from the mapped processes into an error log entry or a normal log entry for the training of the first machine learning algorithm.
20. The non-transitory computer readable medium according to claim 16 , wherein the predicted anomalies in the features of the one or more components of the customer-premises equipment using the second machine learning algorithm comprises a defect in code of one or more of the components of the customer-premises equipment.
21. The method according to claim 1 , wherein the executing the fix to code of the customer-premises equipment further comprises:
transmitting, by the processing server, the fix to the customer-premises equipment.
22. The non-transitory computer readable medium according to claim 16 , wherein the executing the fix to code of the customer-premises equipment further comprises:
transmitting, by the processing server, the fix to the customer-premises equipment.