For application in the PCR model, the T matrix was split in parts relating to the reference and test datasets (T r and T t, respectively), using methods that are briefly described in Additional file 1: Figure S1.
This variance-covariance matrix was split up to resemble the separate random cow-effects (K i.G, B i.G and C i.G ) and the random "environmental" effects (K i.E, B i.E and C i.E ), such that simulated lactation curves yielded realistic curves.
By the dot layer, the InGaAs matrix is split into two parts.
The M × N image matrix is split up into 2 × 2 overlapped blocks as shown in Fig. 4.
Hybrid methods require that the make matrix is split into two matrices, V 1 and V 2, where in this case the first one includes outputs for which the commodity technology assumption is made and the second includes those, which are to be treated on an industry technology assumption.
The matrix is split in two parts: the first T lines represent the training set, while the remaining L= N− T lines represent the testing set.
After the modal analysis of a non-gyroscopic undamped system having the same inertial and elastic properties of the rotor has been performed, the damping and gyroscopic matrices are split into a "proportional" and a "non-proportional" part.
Taking into account the frequency-selective feature of the channel, the simulated MIMO channel matrices are split into Q = 512 subchannels, i.e., each subchannel has a bandwidth of 39.0625 kHz, which corresponds to the coherence bandwidth of channels that have a delay spread of 12.8 μs[3].
Without going into details, this depends on (i) the way matrices are split between parallel tasks, (ii) the low number of interface nodes (i.e., nodes shared by different tasks) typical of SEM, as they need less grid points to provide the same accuracy, and (iii) the fact that parallel efficiency roughly depends on the ratio between computational effort of each task and amount of message passing.
In addition, they found that the best SMQ performance at low rates was achieved when the spectral parameters matrix (assuming a size for each matrix of LSFs) was split into five equal dimension size submatrices, =, given by (1).
For every tag, the sum of normalized frequencies was calculated and tags were discarded for values less than 2. The resulting matrix (2,918 rows) was split into two parts: the first registering the 500 tags with the highest sum of frequencies (Top500) and the second registering tags with lower frequencies (2,418).
