U.S. flag

An official website of the United States government, Department of Justice.

Enhanced Ignitable Liquid and Substrate Database Functionality for Improved Casework and Research.

Award Information

Award #
Funding Category
Competitive Discretionary
Congressional District
Funding First Awarded
Total funding (to date)

Description of original award (Fiscal Year 2022, $289,496)

The Ignitable Liquids Reference Collection (ILRC) Database is over 20 years old, and the Substrate Database is 14 years old. The two databases work independently and cannot be searched jointly. Information from the two databases cannot be viewed together adding complexity for data analysis and interpretation. From the casework perspective, the information contained in the databases is underutilized due to the current database structure. It is proposed to add functions such as digital weathering to the igntiable liquids, multiple compound search, and selecting the ratio of ignitable liquid contribution to that of substrate for creating insilico fire debris. Visualization of the insilico fire debris to a casework sample will aid the analyst in interpretation and being able to illustrate the comparison. Research has also shown that it is possible to digitally combine records from the ILRC and Substrate databases to generate training data sets for machine learning. It is also possible to select casework relevant populations from the database. Neither of these features are supported by the current database structure, yet from a research perspective these are important features.  Data records from the existing databases will be incorporated into the enhanced database that will be designed for both casework and research needs. Many of the proposed functions have been developed in research already and will be incorporated into the database. It is proposed to program a new interface that supports the new features discussed above and test the improved database functionality. In addition, the database will be tested by creating a custom dataset and utilizing it in a machine learning classifier. Outcomes from this work will include a prototype database for fire debris analysis that will be valuable for both casework and research. Implementation and testing with a machine learning methodology will be demonstrated. CA/NCF

Date Created: September 28, 2022