Abstract : The Large Scale Classification Project at Camp Butner provides an excellent opportunity to compare and contrast classification performances for static and reconnaissance EMI data and for a variety of analysis approaches. SAIC analyzed EM61 data acquired in reconnaissance mode as well as Metal Mapper and TEMTADS data acquired while stationary. Our analysis included single- and multi-source solvers. Our classification utilizes a decision tree targeting the intrinsic polarizabilities. The decision tree incorporates uncertainty in unanticipated targets-of-interest and has hasn't changed dramatically since being developed using data acquired at Aberdeen Proving Ground, Camp Sibert, and Camp San Luis Obispo. We also experimented in the number of training labels (starting with no on-site labels) used to fine tune the classifier. Finally, we utilized two different analysis environments; Oasis montaj and IDL. Two commercial firms, NAEVA and Parsons, also utilized the UX-Analyze module in Oasis montaj to classify Metal Mapper stationary data. During our presentation, we will discuss performances of the various combinations and present lessons learned.