Division Deliverables

• For each of the three approaches above, experimental automatic processing pipelines will be setup, integrating the three algorithms under consideration (each in a separate pipeline). This implies automation of some manual steps in these methods.

• Each pipeline will run on the same set of representative events that cause aftershocks.

• A set of automatic tests will be designed and implemented allowing collecting statistical information on the three algorithms when run on a representative set of events, for performance comparison purposes.

• Statistical data collected as a result of automatic tests will be used to compare the performance of the algorithms on representative data sets.

• SHI MDA analysts will also assess the results produced by the three algorithms from the point of view of their quality as a starting point for analyst review.

• The final deliverables should be a report and recommendation summarising the findings above establishing which of the three approaches (if any) should be pursued for further development and implementation in an operational system. This should include estimates for the remaining effort needed to complete development.

The project will be executed over a duration of 1,5 years, starting in the second quarter of 2018. It is estimated that about 60 % of the total effort, mainly in the first year of the project, will be spent on setting up the experimental pipelines. The remainder of the total effort will be dedicated to designing automatic tests, collecting their results and analysing results.