Determining a risk indicator based on classifying documents using a classifier
Systems and techniques for determining and displaying risk indicators are described. A set of event logs occurring in a time interval and associated with a user account retrieving may be retrieved from an event log database. For individual event logs in the set of event logs, a context may be determined. A resource associated with the individual log may have an associated classification. An activity risk associated with the individual event log may be determined based at least in part on the first context and, if applicable, on the associated classification. For individual event logs in the set of event logs, a risk indicator may be determined based at least in part on the activity risk. In some cases, a cumulative risk indicator may be determined for the particular time interval based on the risk indicators associated with the individual event logs.
1. A computer-implemented method, comprising:
generating a classifier based on a classifier algorithm;
training, using electronic training data, the classifier to classify data within a predetermined accuracy, wherein training the classifier to classify data within the predetermined accuracy comprises:
training the classifier with training data;
classifying, by the classifier, test data to generate classified test data;
determining that the test data is not within the predetermined accuracy based on a comparison of the classified test data with a pre-classified set of test data;
tuning the classifier;
retraining the classifier with the training data;
reclassifying, by the classifier, the test data to generate reclassified test data;
determining that the reclassified test data is within the predetermined accuracy based on a comparison of the reclassified test data with the pre-classified set of test data;
receiving, via a network, a first event log from an agent that is configured to monitor data activities at a data storage, the first event log including information pertaining to a first access to data stored in the data storage by an electronic device associated with a user account;
causing the first event log to be stored in an event log database;
retrieving, from the event log database, a set of event logs associated with a user account, each event log of the set of event logs occurring in a time interval;
selecting the first event log from the set of event logs, the first event log associated with a resource having a classification assigned to the resource by the classifier based at least in part on a taxonomy;
determining a first context associated with the first event log;
determining a first activity risk associated with the first event log based at least in part on the first context;
determining a first risk indicator associated with the user account based at least in part on the first activity risk;
receiving, via the network, a second event log from the agent during the time interval, the second event log including information pertaining to a second access to data stored in the data storage by the electronic device associated with the user account;
determining a second risk indicator associated with the user account based at least in part on the second event log; and
determining a cumulative risk indicator for the time interval based on the first risk indicator and the second risk indicator.
2. The computer-implemented method of claim 1 , further comprising:
selecting a third event log from the set of event logs;
determining a second context associated with the third event log;
determining a second activity risk associated with the third event log based at least in part on the second context; and
determining a third risk indicator associated with the user account based at least in part on the second activity risk.
3. The computer-implemented method of claim 2 , further comprising:
displaying, for the time interval, a graph that includes a plurality of risk indicators including at least the first risk indicator the second risk indicator, and the third risk indicator.
4. The computer-implemented method of claim 2 , further comprising:
determining a first cumulative risk indicator associated with the user account during a first time interval based at least in part on the first risk indicator and the second risk indicator;
determining a second cumulative risk indicator associated with the user account during a second time interval based at least in part on a second set of event logs; and
displaying a graph that includes a plurality of cumulative risk indicators including at least the first cumulative risk indicator and the second cumulative risk indicator.
5. The computer-implemented method of claim 1 , further comprising:
determining the resource identified in the first event log as being accessed by an activity;
determining the classification associated with the resource; and
associating a resource risk associated with the resource based at least in part on the classification.
6. The computer-implemented method of claim 5 , determining the first activity risk associated with the first event log based at least in part on the first context comprises:
determining the first activity risk based at least in part on the first context and the resource risk associated with the resource.
7. The computer-implemented method of claim 5 , wherein:
the classification associated with the resource comprises one of public, internal, confidential, or restricted.
8. One or more non-transitory computer-readable media storing instructions that are executable by one or more processors to perform operations comprising:
generate a classifier based on a classifier algorithm;
train, using electronic training data, the classifier to classify data within a predetermined accuracy, wherein the operation to train the classifier to classify data within the predetermined accuracy further comprises:
train the classifier with training data;
classify, by the classifier, test data to generate classified test data;
determine that the test data is not within the predetermined accuracy based on a comparison of the classified test data with a pre-classified set of test data;
tune the classifier;
retrain the classifier with the training data;
reclassify, by the classifier, the test data to generate reclassified test data;
determine that the reclassified test data is within the predetermined accuracy based on a comparison of the reclassified test data with the pre-classified set of test data;
receive, via a network, a first event log from an agent that is configured to monitor data activities at a data storage, the first event log including information pertaining to a first access to data stored in the data storage by an electronic device associated with a user account;
cause the first event log to be stored in an event log database;
retrieve, from the event log database, a set of event logs associated with a user account, each event log of the set of event logs occurring in a time interval;
select the first event log from the set of event logs, the first event log associated with a resource having a classification assigned to the resource by the classifier based at least in part on a taxonomy;
determine a first context associated with the first event log;
determine a first activity risk associated with the first event log based at least in part on the first context;
determine a first risk indicator associated with the user account based at least in part on the first activity risk;
receive, via the network, a second event log from the agent during the time interval, the second event log including information pertaining to a second access to data stored in the data storage by the electronic device associated with the user account;
determine a second risk indicator associated with the user account based at least in part on the second event log; and
determine a cumulative risk indicator for the time interval based on the first risk indicator and the second risk indicator.
9. The one or more non-transitory computer-readable media of claim 8 , the operations further comprising:
select a third event log from the set of event logs;
determine a second context associated with the third event log;
determine a second activity risk associated with the third event log based at least in part on the second context; and
determine a third risk indicator associated with the user account based at least in part on the second activity risk.
10. The one or more non-transitory computer-readable media of claim 9 , the operations further comprising:
display, for the time interval, a graph that includes a plurality of risk indicators including at least the first risk indicator, the second risk indicator, and the third risk indicator.
11. The one or more non-transitory computer-readable media of claim 9 , the operations further comprising:
determine a first cumulative risk indicator associated with the user account during a first time interval based at least in part on the first risk indicator and the second risk indicator;
determine a second cumulative risk indicator associated with the user account during a second time interval based at least in part on a second set of event logs; and
display a graph that includes a plurality of cumulative risk indicators including at least the first cumulative risk indicator and the second cumulative risk indicator.
12. The one or more non-transitory computer-readable media of claim 8 , the operations further comprising:
determine the resource identified in the first event log as being accessed by an activity;
determine the classification associated with the resource; and
associate a resource risk associated with the resource based at least in part on the classification.
13. The one or more non-transitory computer-readable media of claim 8 , determining the first activity risk associated with the first event log based at least in part on the first context comprises:
determine the first activity risk based at least in part on the first context and a resource risk associated with the resource, wherein the classification associated with the resource comprises one of public, internal, confidential, or restricted.
14. A server, comprising:
one or more processors; and
one or more non-transitory computer-readable media storing instructions that are executable by the one or more processors to perform operations comprising:
generate a classifier based on a classifier algorithm;
train, using electronic training data, the classifier to classify data within a predetermined accuracy, wherein the operation to train the classifier to classify data within the predetermined accuracy further comprises:
train the classifier with training data;
classify, by the classifier, test data to generate classified test data;
determine that the test data is not within the predetermined accuracy based on a comparison of the classified test data with a pre-classified set of test data;
tune the classifier;
retrain the classifier with the training data;
reclassify, by the classifier, the test data to generate reclassified test data;
determine that the reclassified test data is within the predetermined accuracy based on a comparison of the reclassified test data with the pre-classified set of test data;
receive, via a network, a first event log from an agent that is configured to monitor data activities at a data storage, the first event log including information pertaining to a first access to data stored in the data storage by an electronic device associated with a user account;
cause the first event log to be stored in an event log database;
retrieve, from the event log database, a set of event logs associated with a user account, each event log of the set of event logs occurring in a time interval;
select the first event log from the set of event logs, the first event log associated with a resource having a classification assigned to the resource by the classifier based at least in part on a taxonomy;
determine a first context associated with the first event log;
determine a first activity risk associated with the first event log based at least in part on the first context;
determine a first risk indicator associated with the user account based at least in part on the first activity risk;
receive, via the network, a second event log from the agent during the time interval, the second event log including information pertaining to a second access to data stored in the data storage by the electronic device associated with the user account;
determine a second risk indicator associated with the user account based at least in part on the second event log; and
determine a cumulative risk indicator for the time interval based on the first risk indicator and the second risk indicator.
15. The server of claim 14 , the operations further comprising:
select a third event log from the set of event logs;
determine a second context associated with the third event log;
determine a second activity risk associated with the third event log based at least in part on the second context; and
determine a third risk indicator associated with the user account based at least in part on the second activity risk.
16. The server of claim 15 , the operations further comprising:
display, for the time interval, a graph that includes a plurality of risk indicators including at least the first risk indicator, the second risk indicator, and the third risk indicator.
17. The server of claim 15 , the operations further comprising:
determine a first cumulative risk indicator associated with the user account during a first time interval based at least in part on the first risk indicator and the second risk indicator;
determine a second cumulative risk indicator associated with the user account during a second time interval based at least in part on a second set of event logs; and
display a graph that includes a plurality of cumulative risk indicators including at least the first cumulative risk indicator and the second cumulative risk indicator.
18. The server of claim 14 , the operations further comprising:
determine the resource identified in the first event log as being accessed by an activity;
determine the classification associated with the resource; and
associate a resource risk associated with the resource based at least in part on the classification.
19. The server of claim 18 , wherein determining the first activity risk associated with the first event log based at least in part on the first context comprises:
determine the first activity risk based at least in part on the first context and the resource risk associated with the resource.
20. The server of claim 18 , wherein:
the classification associated with the resource comprises one of public, internal, confidential, or restricted.