Reinforcement learning approach to root cause analysis

1. A computer-implemented method comprising:

generating, by a processor, a vector representation of a root cause of an error based on a hierarchical topology of a computing system;

generating, by the processor, a respective vector representation of each subject matter expert of a plurality of subject matter experts based at least in part on the hierarchical topology of the computing system, wherein the hierarchical topology of the computing system is a tree-type data structure;

selecting, by the processor, a subject matter expert from the plurality of subject matter experts based at least in part on the vector representation of the root cause of the error; and

uploading, by the processor, a diagnostic software to the computing system,

wherein selecting the subject matter expert from the plurality of subject matter experts comprises:

assigning rankings to the respective vector representation of each subject matter expert of the plurality of subject matter experts;

inputting, to a machine learning algorithm, an assigned vector representation of a root cause of an error and the respective vector representation of each subject matter expert of the plurality of subject matter experts and a corresponding ranking of the respective vector representation of each subject matter expert of the plurality of subject matter experts; and

training the machine learning algorithm to select the subject matter expert from the plurality of subject matter experts corresponding to the assigned vector representation of the root cause of the error by rewarding the machine learning algorithm in response to a given subject matter expert among the plurality of subject matter experts resolves the assigned vector representation of the root cause of the error, and penalizing the machine learning algorithm in response to a given subject matter expert among the plurality of subject matter experts failing to resolve the assigned vector representation of the root cause of the error.

2. The computer-implemented method of claim 1 , further comprising:

monitoring a progress of the selected subject matter expert from the plurality of subject matter experts in resolving the error; and

modifying the assigned vector representation of the root cause of the error based at least in part on whether the selected subject matter expert from the plurality of subject matter experts resolved the error.

3. The computer-implemented method of claim 2 , further comprising adjusting a ranking of the selected subject matter expert from the plurality of subject matter experts based at least in part on the progress of the selected subject matter expert from the plurality of subject matter experts in resolving the error.

4. The computer-implemented method of claim 1 , wherein the error is associated with a node of the tree-type data structure.

5. The computer-implemented method of claim 1 , further comprising:

calculating a weight of each anomaly detected;

calculating an anomaly score for each leaf node in the tree-type data structure, wherein the anomaly score is based at least in part on the weight of each anomaly detected;

calculating an anomaly score for parent node of each leaf node by summing together a respective anomaly score of each leaf node that is a child node of the parent node of each leaf node, and

determining the root cause of the error based at least in part on a respective anomaly score of each node in the tree-type data structure.

6. The computer-implemented method of claim 5 , further comprising:

collecting time series data at each node in the tree-type data structure;

comparing the collected time series data with respective forecasted time series data to detect a deviation from the respective forecasted time series data; and

calculating the anomaly score based on a degree of deviation from the respective forecasted time series data.

7. A system comprising:

a memory storing computer readable instructions; and

one or more processors executing the computer readable instructions, the computer readable instructions controlling the one or more processors to perform operations comprising:

generating a vector representation of a root cause of an error based on a hierarchical topology of a computing system;

generating a respective vector representation of each subject matter expert of a plurality of subject matter experts based at least in part on the hierarchical topology of the computing system, wherein the hierarchical topology of the computing system is a tree-type data structure;

selecting a subject matter expert from the plurality of subject matter experts based at least in part on the vector representation of the root cause of the error; and

uploading a diagnostic software to the computing system,

wherein selecting the subject matter expert from the plurality of subject matter experts comprises:

assigning rankings to the respective vector representation of each subject matter expert of the plurality of subject matter experts;

inputting, to a machine learning algorithm, an assigned vector representation of a root cause of an error and the respective vector representation of each subject matter expert of the plurality of subject matter experts and a corresponding ranking of the respective vector representation of each subject matter expert of the plurality of subject matter experts; and

training the machine learning algorithm to select the subject matter expert from the plurality of subject matter experts corresponding to the assigned vector representation of the root cause of the error by rewarding the machine learning algorithm in response to a given subject matter expert among the plurality of subject matter experts resolves the assigned vector representation of the root cause of the error, and penalizing the machine learning algorithm in response to a given subject matter expert among the plurality of subject matter experts failing to resolve the assigned vector representation of the root cause of the error.

8. The system of claim 7 , wherein the operations performed by the one or more processors further comprise:

monitoring a progress of the selected subject matter expert from the plurality of subject matter experts in resolving the error; and

modifying the assigned vector representation of the root cause of the error based at least in part on whether the selected subject matter expert from the plurality of subject matter experts resolved the error.

9. The system of claim 8 , wherein the operations performed by the one or more processors further comprise adjusting a ranking of the selected subject matter expert from the plurality of subject matter experts based at least in part on the progress of the selected subject matter expert from the plurality of subject matter experts in resolving the error.

10. The system of claim 7 , wherein the error is associated with a node of the tree-type data structure.

11. The system of claim 7 , wherein the operations performed by the one or more processors further comprise:

calculating a weight of each anomaly detected;

calculating an anomaly score for each leaf node in the tree-type data structure, wherein the anomaly score is based at least in part on the weight of each anomaly detected;

calculating an anomaly score for parent node of each leaf node by summing together a respective anomaly score of each leaf node that is a child node of the parent node of each leaf node, and

determining the root cause of the error based at least in part on a respective anomaly score of each node in the tree-type data structure.

12. The system of claim 11 , wherein the operations performed by the one or more processors further comprise:

collecting time series data at each node in the tree-type data structure;

comparing the collected time series data with respective forecasted time series data to detect a deviation from the respective forecasted time series data; and

calculating the anomaly score based on a degree of deviation from the respective forecasted time series data.

13. A computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to perform operations comprising:

generating a vector representation of a root cause of an error based on a hierarchical topology of a computing system;

generating a respective vector representation of each subject matter expert of a plurality of subject matter experts based at least in part on the hierarchical topology of the computing system, wherein the hierarchical topology of the computing system is a tree-type data structure;

selecting a subject matter expert from the plurality of subject matter experts based at least in part on the vector representation of the root cause of the error; and

uploading a diagnostic software to the computing system,

wherein selecting the subject matter expert from the plurality of subject matter experts comprises:

assigning rankings to the respective vector representation of each subject matter expert of the plurality of subject matter experts;

inputting, to a machine learning algorithm, an assigned vector representation of a root cause of an error and the respective vector representation of each subject matter expert of the plurality of subject matter experts and a corresponding ranking of the respective vector representation of each subject matter expert of the plurality of subject matter experts; and

training the machine learning algorithm to select the subject matter expert from the plurality of subject matter experts corresponding to the assigned vector representation of the root cause of the error by rewarding the machine learning algorithm in response to a given subject matter expert among the plurality of subject matter experts resolves the assigned vector representation of the root cause of the error, and penalizing the machine learning algorithm in response to a given subject matter expert among the plurality of subject matter experts failing to resolve the assigned vector representation of the root cause of the error.

14. The computer program product of claim 13 , wherein the operations performed by the processor further comprise:

monitoring a progress of the selected subject matter expert from the plurality of subject matter experts in resolving the error; and

modifying the assigned vector representation of the root cause of the error based at least in part on whether the selected subject matter expert from the plurality of subject matter experts resolved the error.

15. The computer program product of claim 13 , wherein the error is associated with a node of the tree-type data structure.

16. The computer program product of claim 13 , wherein the operations performed by the processor further comprise:

calculating a weight of each anomaly detected;

calculating an anomaly score for each leaf node in the tree-type data structure, wherein the anomaly score is based at least in part on the weight of each anomaly detected;

calculating an anomaly score for parent node of each leaf node by summing together a respective anomaly score of each leaf node that is a child node of the parent node of each leaf node; and

determining the root cause of the error based at least in part on a respective anomaly score of each node in the tree-type data structure.

17. The computer program product of claim 16 , wherein the operations performed by the processor further comprise:

collecting time series data at each node in the tree-type data structure;

comparing the collected time series data with respective forecasted time series data to detect a deviation from the respective forecasted time series data; and

calculating the anomaly score based on a degree of deviation from the respective forecasted time series data.