Advanced Data Management and Mining for BMDS Digital Simulations

Sherlock provides a data mining solution that automates the discovery of correlations and causations between BMDS performance and critical factors within large amounts of missile defense simulation data. The future missile defense simulations will include digital simulations capable of generating orders of magnitude more data than can be handled by the current MDA Extract, Transform, and Load (ETL) and analysis tools within MDA. Sherlock enables the effective analysis and assessment of this big data. Sherlock is composed of a flexible ETL platform and a tailored, state of the art data mining platform. The ETL platform enables the data mining platform to dynamically request new data transformations (views and schemas) to prepare the data in the format needed for the data mining platform. The data mining platform includes machine learning algorithms that continually explore the test space to identify end-to-end performance drivers for the analyst and assessor. Sherlock will allow tailored access to data while retaining standard internal data structures and associations, as well as providing a core set of tools for all users. Sherlock is needed to fully exploit the future MDA digital simulations data so that MDA can make comprehensive assessments of the new BMDS capabilities. Approved for Public Release | 18-MDA-9817 (23 Oct 18)