Therefore GAU2 is an efficient measure to help decision maker to evaluate more accurate results and smoother decisions are made in final decisions by DSmT in comparison to DST.
Next, we re-tested classification performance with the new smoothed decision rules that should contain more meaningful features.
Similar to the case of DFA classifiers, the performance here (for the smoothed decision echoes) was not as good as that achieved by using a dolphin signal.
Moreover, the high performance of the smoothed decision echoes (see Results) implies that the temporal resolution can be reduced without harming classification much.
The smoothed decision echoes however (see Materials and Methods) still enable highly significant classification performance (t-test, Table 3) and reveal salient structures that could imply a meaningful decision rules.
The smoothed decision echoes (Figure 2) depicting the rules learned by the classifier, are characterized by clear hyper and hypo-intensity blob-like structures that represent areas in the spectral-temporal representation of the echoes that are important for classification.
These rules, as appear in the smoothed decision echoes, seem to contain both temporal and spectral features which cover the entire spectral-temporal range of the echoes (Figure 2B&C).
The results for the smoothed decision echoes are given in Tables 3 and 4. The Results for the noise sensitivity test are given in Tables 5 and 6 and the performance for the principle component analysis resampled data were still high above change level in most cases (last row in Tables 5 and 6).
As in the dolphin case, the smoothed decision echoes of these classifiers perform very well (t-test, Table 4) and show clear structures that seem to contain both temporal and spectral cues which cover the entire spectral-temporal range of the echo (Figure 2D&E).
Mutually suspicious relations between the prime minister's aides, government bureaucrats and company officials obstructed smooth decision-making.
