"... In PAGE 6: ... Some stages may take several processor cycles. Table1 details the characteristics of the proces- sors used in this work: a 4-way processor (resembles the Alpha 21264 processor) and a 8-way processor. In our evaluations we used oracle memory disam- biguation.... ..."

"... In PAGE 7: ... Parame- ters of the GA are given in Table 2. The algorithm was run 5 times and the results ob- tained using real-world data for one month planning horizon discretised homogene- ously into one minute unit time periods, are given in Table3 . This specific month was chosen because it was considered to be rather busy with 158 jobs and a workload of 5120125 required items.... In PAGE 7: ... Parameters of the genetic algorithm for load balancing Population size 50 Length of the chromosome Two layers: The first layer has the number of jobs to be processed in one work-centre The second layer has the number of lots to be processed in the work-centre Mutation probability 0.4 Termination condition 500 iterations Table3 presents best values of the fitness function (the average satisfaction grades of all objectives) of the genetic algorithm for scheduling using 4 different ways of machines loading and the corresponding values of the satisfaction grades of the objec- tive functions given in (1)-(5) together with the makespan Cmax which is the time when the last job finishes its processing. Table3 Performance of schedules using 4 different ways of machine loading ... ..."

"... In PAGE 11: ...In Table1 , I have compared the classifications of studies that different authors have proposed. ... In PAGE 12: ... What seems a little questionable, is their choice of labels for these categories. Therefore, in the first row of Table1 , I have proposed new labels and short definitions for the three broad categories of research designs. For better and more detailed descriptions of different possibilities of combining quantitative and qualitative approaches, this broad classification needs to be divided into subcategories.... ..."

"... In PAGE 27: ...23 Two-Way Communication Tools and Tiered Rank Results indicate, however, that use of two-way tools on the entire sample of Fortune 500 corporate web pages did not differ statistically by rank. As Table4 illustrates, none of the five two-way tool categories produced significantly different results by tiered rank. Overall two-way tool presence was especially minimal, and in some cases, non-existent.... ..."

"... In PAGE 4: ...alues over the 10 clusters analyzed by Eisen et al. [8] is shown in Figure 3. The significance of the differences between these clusters in terms of their GO-driven similarity was established by a one-way ANOVA. The results shown in Table2 confirm that the functional differences between these clusters are significant (p lt; 0.0005).... ..."

"... In PAGE 6: ... Then these selected data are used as inputs to NTF, in order to learn compo- nent matrices A(n). In some cases where the data tensor involves both temporal and trial coordinates, we use a whole data set without the data selection step, since it destroys the structure (for example, the case of X 2 R27 2 Table2 K in Table 2). We use the nearest neighbor method for data se- lection.... In PAGE 8: ... The other di erence between data set is the classi er. The Graz dataset use the Gaussian probabilistic model combining the information across time (except for the 4-way tensor (time) structure in Table2 , because the temporal information can be explained by the the component matrix A(3) which involved in the temporal dimension, in this case, we just use the Gaussian probabilistic model only taking a single time point into account), while the IDIAP dataset use the Viterbi algorithm because it has no trial structure, thus, there is no xed tendency across time. For Graz dataset, NMF is enough to extract the meaningful features which discriminate between classes, thus, the extension to NTF is not necessary.... In PAGE 9: ... 6. they are obtained by NMF, 3-way NTF, and NTF with 4-way tensor (class) in Table2 (from top to bottom) and the number of basis r = 2; 4; 5; 6 (from left to right). In each plot, top 1/2 is associated with C3 and bottom 1/2 is contributed by C4.... ..."