maximum likelihood linear regression.
Maximum likelihood linear regression (MLLR) [36] was employed for adaptation.
Maximum likelihood linear regression (MLLR) [33, 42], maximum a posteriori (MAP) [34, 24], parallel model combination (PMC) [23, 31, 33] and a novel enhanced version of PMC, PCA and CMS based PMC (PC-PMC) [30] are well incorporated in the system.
Unfortunately, the Guan et al. ([2006]) approach used to estimate production risks and the bivariate probit model used to estimate adoption are non-linear maximum likelihood models which cannot be directly estimated by fixed effects (Wooldridge, [2002]).
Similar to maximum normalized likelihood linear filtering in [45], the parameters of the cross transform can be estimated by minimizing an approximated KL divergence between the distribution of the processed features, p y, and the distribution of clean training features, p Λ.
FPKM is an estimate of transcriptional abundance estimated by the approximately linear maximum abundance likelihood model, taking into account stochastic sequencing biases, described in detail by Trapnell et al. (2010).
Some examples of the model adaptation method include the (constrained) maximum likelihood linear regression (CMLLR/MLLR) [30], maximum a posteriori (MAP) [37], and vector Taylor series (VTS) [38 40] based adaptation methods.
This latter method exploits an invaluable prior knowledge attained from an average voice model [3], and adapts this general model using an adaptation algorithm such as maximum likelihood linear regression (MLLR)[32], maximum a posteriori (MAP)[33], and cluster adaptive training (CAT [21].
This is in spirit similar to the popular feature space maximum likelihood linear regression (fMLLR) [30] that uses the maximum likelihood criterion to estimate the linear transform.
Some examples of backend approaches are Parallel Model Combination (PMC) [5] and model adaption algorithms, such as Maximum Likelihood Linear Regression (MLLR) [6] and Maximum A Posterior probability (MAP) based adaptation [7].
