This directory contains results for 8-component MAR with a simplified (silence removed) WSJ.
(containing 20 speakers, and 6 additional speakers for bgd model).

/data/isip002/research/nonlinear/mix_ar/mar/exp011_mix8_order1

The speaker recognition result is in:

http://www.ece.msstate.edu/research/isip/projects/nsf_nonlinear/doc/baseline_results/speaker_recognition/mix_ar/mar/exp_011/mar_orders_0_1.jpg

The MAR training and likelihood calculation for order 1 was calculated for a threshold of 0.01. (threshold: absolute difference in MFCC values for 2 successive frames).

The DET plots are somewhat similar, and that for order 1 is marginally better.