Abstract by Makenzie Allen
David Van Komen
Physics and Astronomy
Investigating parameter importance for different ocean environments using Fisher Information
Physical models of underwater acoustic propagation often involve many unknown parameters. Overly complex models may be difficult to calibrate and make inaccurate predictions. One way to find an effective model is to consider the relative impact of parameters on predictions as seen in the Fisher Information Matrix (FIM). I demonstrate the applicability of this approach using a model of transmission loss for different ocean environments. The eigenvalues of the FIM shows that the system is “sloppy”, i.e., it has an exponential hierarchy of parameter importance. In many cases, only a few parameters are relevant for explaining the model output. However, the relevance of individual physical parameters vary with both environment and frequency. These results have implications for learning algorithms and data collection methods while elucidating the relevant physics for different conditions.