Automatic classification of correlated anomalies from a network through interpretable clustering
A method includes receiving network data describing operation of a network including a plurality of anomalies; clustering the network data to obtain clusters of groups of correlated anomalies; responsive to labeling the clusters of groups of correlated anomalies, utilizing the labels for the network data to train a model for automatic classification; and providing the model for automatic classification of additional network data. The clusters can be described through interpretable clustering (tree learning on cluster labels) (usage for description) or the cluster of interest can be deployed (again, tree learning on cluster labels) (usage for automatic classification).
1. A method comprising steps of:
receiving network data describing operation of a network including a plurality of groups of anomalies;
clustering the network data to obtain a plurality of clusters, each of the plurality of clusters comprising one or more groups of correlated anomalies;
responsive to labeling the clusters of groups of correlated anomalies, utilizing the labels for the network data to train a model for predicting cluster labels for incoming network data;
utilizing the model in production for predicting cluster labels for incoming network data, thereby determining a category of new incoming groups of correlated anomalies, wherein the predicting is based on data characteristics associated with the network data, the data characteristics including presence of approximate periodicities,
utilizing the model to determine a category of new incoming groups of correlated anomalies in the real time; and
employing a new model in place of the model and utilizing the new model to determine a category of new incoming groups of correlated anomalies in real time without loss of real time classification information.
2. The method as claimed in claim 1 , wherein the labeling is performed by a human expert that determines the labels, and wherein each label is a naming of each cluster, the naming being in a form that is comprehensible to a human.
3. The method as claimed in claim 1 , wherein the labeling is performed automatically utilizing a decision tree that provides a cluster description for the label.
4. The method as claimed in claim 1 , wherein the steps further include
determining the clusters of groups of correlated anomalies requires additional data and repeating the receiving and clustering.
5. The method as claimed in claim 1 , wherein the clustering includes unsupervised machine learning on the network data and the utilizing includes supervised machine learning.
6. The method as claimed in claim 1 , wherein the steps further include
evaluating an accuracy of the model; and
selecting a level of accuracy based on the evaluating.
7. The method as claimed in claim 1 , wherein the network data includes at least one metric/dimension linked to each of the plurality of groups of correlated anomalies, the at least one metric/dimension being selected from: Group duration, Group impacted subscribers, Severity, User Flags, Number of anomalies contained in group, Root cause diagnosis, Presence of approximate periodicities, Nature of rules linked to anomalies, Nature of root cause diagnosis elements targeted, Nature of root cause diagnosis 3GPP cause and Geographic data linked to dimensions, and wherein the at least one metric/dimension is utilized for performing the clustering.
8. The method as claimed in claim 1 , wherein the network is a wireless network.
9. The method as claimed in claim 1 , wherein the clustering utilizes any of K-means, agglomerative clustering, DBSCAN.
10. The method as claimed in claim 9 , wherein a required number of clusters is automatically determined using the elbow method.
11. The method as claimed in claim 1 , wherein the steps further include
providing a user interface that allows a user to select a profile of the groups of correlated anomalies for data analysis.
12. The method as claimed in claim 11 , wherein the steps further include
providing a label based on the profile.
13. A server comprising:
at least one processor and memory comprising instructions that, when executed, cause the at least one processor to:
receive network data describing operation of a network including a plurality of groups of anomalies;
cluster the network data to obtain a plurality of clusters, each of the plurality of clusters comprising one or more groups of correlated anomalies;
responsive to labeling the clusters of groups of correlated anomalies, utilize the labels for the network data to train a model for predicting cluster labels for incoming network data;
utilize the model in production for predicting cluster labels for incoming network data, thereby determining a category of new incoming groups of correlated anomalies, wherein the predicting is based on data characteristics associated with the network data, the data characteristics including presence of approximate periodicities;
utilizing the model to determine a category of new incoming groups of correlated anomalies in the real time; and
employing a new model in place of the model and utilizing the new model to determine a category of new incoming groups of correlated anomalies in real time without loss of real time classification information.
14. The server as claimed in claim 13 , wherein the labeling is one or more of
performed by a human expert that determines the labels, wherein each label is a naming of each cluster, the naming being in a form that is comprehensible to a human; and
performed automatically utilizing a decision tree that provides a cluster description for the label.
15. The server as claimed in claim 13 , wherein the instructions that, when executed, cause the at least one processor to:
determine the clusters of groups of correlated anomalies requires additional data and repeating the receiving and clustering.
16. The server as claimed in claim 13 , wherein the clustering includes unsupervised machine learning on the network data and the utilizing includes supervised machine learning.
17. A non-transitory computer-readable medium having instructions stored thereon for programming at least one processor to perform steps of:
receiving network data describing operation of a network including a plurality of groups of anomalies;
clustering the network data to obtain a plurality of clusters, each of the plurality of clusters comprising one or more groups of correlated anomalies;
responsive to labeling the clusters of groups of correlated anomalies, utilizing the labels for the network data to train a model for predicting cluster labels for incoming network data;
utilizing the model in production for predicting cluster labels for incoming network data, thereby determining a category of new incoming groups of correlated anomalies, wherein the predicting is based on data characteristics associated with the network data, the data characteristics including presence of approximate periodicities;
utilizing the model to determine a category of new incoming groups of correlated anomalies in the real time; and
employing a new model in place of the model and utilizing the new model to determine a category of new incoming groups of correlated anomalies in real time without loss of real time classification information.
18. The non-transitory computer-readable medium as claimed in claim 17 , wherein the labeling is one or more of
performed by a human expert that determines the labels, wherein each label is a naming of each cluster, the naming being in a form that is comprehensible to a human; and
performed automatically utilizing a decision tree that provides a cluster description for the label.
19. The non-transitory computer-readable medium as claimed in claim 17 , wherein the clustering includes unsupervised machine learning on the network data and the utilizing includes supervised machine learning.