Multivariate pattern classification, as described in Klöppel et al. [10], was performed with a linear kernel by identifying a separating hyperplane that maximizes the distance between different clinical groups based on whole-brain or ROIs information.
Information from the three single best performing markers and all six markers were combined using the method reported by by Xiong and coworkers [ 27], that is, by identifying the linear combination of markers that yielded the greatest area under the receiver operating characteristic (ROC) curve (AUC).
Frequently, this task is facilitated by identifying linear equality constraints that all integer solutions must satisfy.
Similar to Jenkins et al. (2012a) completeness was identified by a linear increase in the cumulative number of eruptions per unit time.
Differentially expressed genes were then identified by a linear model (Wettenhall and Smyth 2004).
Chondrolabral damage was identified by a linear band of high signal intensity, detected in the labrum or at the chondrolabral transition zone.
The iterative process is terminated when convergence is achieved, which in turn is identified by a linear regression ratio tregress over the minimal scores in the last iterations.
By identifying matrices with corresponding linear operators, [35, Lemma 3.3] gives begin{aligned} {text {Op}}(a) = D_{phi _1}circ A circ C_{phi _2}.
Decoding of visual stimuli by identified linear filters has previously been considered in [42].
Meta-regression models constructed using stratum-specific estimates of effect for age groups, by 10-year age increments, identified a log-linear trend in the effect of UV radiation on disease risk by age stratum, with strongest effects in youngest strata, and effects diminishing with increasing age (change in natural logarithm of IRR = 0.13 per decade, 95% CI 0.04 to 0.23, P = 0.007).
