### Table 5. Regression on Pumadyn with 32 input non-linear with high noise. Training size 64 128 256 512 1024

### Table 3 RSM Linear Regression Data Order Number of

"... In PAGE 8: ...able 2 SLP problem formulation.............................................................................................29 Table3 RSM Linear Regression Data.... In PAGE 47: ... Table3 details the statistical information of the least squares fit to each polynomial. The sixth order polynomial was chosen due to its greater degree of accuracy (particularly its better estimated variance) when compared to the other three response surface models.... ..."

### Table 4: Linear regressions: full 1-day data sets

1998

"... In PAGE 9: ... We then computed linear regressions for each set of samples. The regression lines are shown in Figure 9; the numeric results are given in Table4 . (User-mode regres- sions are given in the table, but not shown in the fig- ure.... ..."

Cited by 97

### Table 4: Linear regressions: full 1-day data sets

1998

"... In PAGE 10: ... We then computed linear regressions for each set of samples. The regression lines are shown in Figure 9; the numeric results are given in Table4 . (User-mode regres- sions are given in the table, but not shown in the fig- ure.... ..."

Cited by 97

### Table 4: Linear regressions: full 1-day data sets

1998

"... In PAGE 9: ... We then computed linear regressions for each set of samples. The regression lines are shown in Figure 9; the numeric results are given in Table4 . (User-mode regres- sions are given in the table, but not shown in the fig- ure.... ..."

Cited by 97

### Table 4. Linear regression analysis Data set Year 2000 Year 2002

"... In PAGE 5: ...90 0.334 276 0.452 In order to provide a baseline, we first ran a linear regression on each data set to examine the straight- line relationships between constructs. The results are shown in Table4 . Note that usefulness, trust, and entertainment are significantly related to intentions in both time periods; response time is not significant in either.... In PAGE 5: ... Table 5 shows, for each of the eight regressions, the added quadratic term, its main effect, and the resulting R2. We note that in Table 5, multicolinearity between the main effect and the quadratic effect has blurred the impact of the main effect to the point that many main effects that were significant in the linear regression in Table4 , are now insignificant. This should not concern us, as the quadratic results for each of these clearly show that the quadratic terms should not be added.... ..."

### Table 4. Resulting r2, q2, and RMSE Values from Fitting Linear or Sigmoidal Models to the Structure-log Pc Dataa

"... In PAGE 4: ...orrelation is slightly improved (r2 ) 0.968, q2 ) 0.942, n ) 8; the dashed line in the Figure 2 insert), indicating a stronger relationship for compounds that are more similar structurally. For compounds B-I, the relation- ships between Pc and PSAd resulting from the different solvation models were not qualitatively different; they appeared rather to be only slightly shifted along the PSAd axis (Figures 2 and 3, Table4 ). When H 216/44 was included in the data set, however, a weaker relationship was obtained between PSAd,chloroform and Pc than between PSAd,vacuum and Pc.... In PAGE 4: ... When H 216/44 was included in the data set, however, a weaker relationship was obtained between PSAd,chloroform and Pc than between PSAd,vacuum and Pc. In contrast, the relationship between PSAd,water and Pc was unaffected by the inclusion of H 216/44 ( Table4 , Figure 3). PSAd and log Dcalc in Predictions of Caco-2 Cell Monolayer Permeability.... In PAGE 7: ...2 ) 0.644, n ) 8tor2 ) 0.485, q2 ) 0.324, n ) 9; Figure 5 and Table4 ). Similar results were obtained by fitting a sigmoidal model to the data (Table 4).... In PAGE 7: ...2 ) 0.644, n ) 8tor2 ) 0.485, q2 ) 0.324, n ) 9; Figure 5 and Table 4). Similar results were obtained by fitting a sigmoidal model to the data ( Table4 ). It is possible that better relationships can be obtained after optimiza- tion of the phospholipid composition of the immobilized liposomes.... ..."

### Table 1) as data vectors.

"... In PAGE 6: ... Table1 : Four Czechs answers What could you conclude about problem of breathing di culties from information you have been given? Most of them were not complete. In fact the only complete one was the information given by your friend from a village.... In PAGE 13: ...0456 0.0000 Table1 0: Application of the conservative modi cation of IPF to the distributions from Table 8 with ordering fP3; P2; P1g. X1 X2 X3 Q0 Q1 Q2 Q3 Q4 .... In PAGE 14: ...Table1 1: Application of the method of iterative averages to the distributions from Table 8 X1 X2 X3 Q0 Q1 Q2 Q3 .... In PAGE 14: ... Example 4 (Application of Method of Iterative Averages) Applying the method of iterative averages to the three two-dimensional distributions fP1(X1; X2); P2(X2; X3); P3(X1; X3)g from Table 8 we get the limit distribution already at 41 steps. As it can be seen from Table1 1, the limit distribution practically does not di er from the results achieved by the conservative modi cation of IPF.... In PAGE 15: ...Table1 2: Comparison of the results achieved with the input system given in Table 8 X1 X2 X3 Qa1 Qa2 Qb Qc Qd 1 1 1 0:0000 0.0000 0.... In PAGE 15: ...0744 0.0711 Table1 3: Description of symbols used in Table 12 Qa1 : : : the barycentre distribution of the limit cycle, when regular IPF procedure is applied to the ordering fP1; P2; P3g Qa2 : : : the barycentre distribution of the limit cycle, when regular IPF procedure is applied to the ordering fP3; P2; P1g Qb : : : the distribution computed with the conservative modi cation of IPF Qc : : : the distribution computed with the method of iterative averages Qd : : : the distribution minimizing the sum of Kullback-Leibler divergences from the input oligodimensional distributions 8 Comparisons of IPF modi cations We have presented two iterative methods converging even for inconsistent systems of oligodimensional distributions. Another distribution with similar properties can be received by application of the regular IPF procedure to the inconsistent system of distributions, and then, if the process starts to cycle, one can simply compute the average (barycentre) of the respective \limit quot; distributions.... In PAGE 16: ...Table1 4: Input distributions fP1; P2; P3; P4g P1 X2 = 1 X2 = 2 P2 X3 = 1 X3 = 2 X1 = 1 0:02 0:38 X2 = 1 0:13 0:17 X1 = 2 0:28 0:32 X2 = 2 0:67 0:03 P3 X4 = 1 X4 = 2 P4 X4 = 1 X4 = 2 X3 = 1 0:1 0:7 X1 = 1 0 + quot; 0:4 ? quot; X3 = 2 0:15 0:05 X1 = 2 0:25 ? quot; 0:35 + quot; Table 15: Comparison of results achieved with input set 4 with quot; = 0:05 (strong consistency) and quot; = 0:2 (weak consistency) X1 X2 X3 X4 Qa ( quot;=0:05) Qb ( quot;=0:05) Qc ( quot;=0:05) Qa ( quot;=0:2) Qb ( quot;=0:2) Qc ( quot;=0:2) 1 1 1 1 0.0009 0.... In PAGE 16: ...P1 X2 = 1 X2 = 2 P2 X3 = 1 X3 = 2 X1 = 1 0:02 0:38 X2 = 1 0:13 0:17 X1 = 2 0:28 0:32 X2 = 2 0:67 0:03 P3 X4 = 1 X4 = 2 P4 X4 = 1 X4 = 2 X3 = 1 0:1 0:7 X1 = 1 0 + quot; 0:4 ? quot; X3 = 2 0:15 0:05 X1 = 2 0:25 ? quot; 0:35 + quot; Table1 5: Comparison of results achieved with input set 4 with quot; = 0:05 (strong consistency) and quot; = 0:2 (weak consistency) X1 X2 X3 X4 Qa ( quot;=0:05) Qb ( quot;=0:05) Qc ( quot;=0:05) Qa ( quot;=0:2) Qb ( quot;=0:2) Qc ( quot;=0:2) 1 1 1 1 0.0009 0.... In PAGE 17: ...Table1 6: Description of used symbols Qa ( quot;=0:05) : : : the result distribution after a regular IPF procedure applica- tion for quot; = 0:05, Qb ( quot;=0:05) : : : the result distribution after an application of the conservative modi cation of IPF for quot; = 0:05, Qc ( quot;=0:05) : : : the result distribution after an application of the method of iterative averages for quot; = 0:05, Qa ( quot;=0:2) : : : the result distribution after a regular IPF procedure applica- tion for quot; = 0:2, Qb ( quot;=0:2) : : : the result distribution after an application of the conservative modi cation of IPF for quot; = 0:2, Qc ( quot;=0:2) : : : the result distribution after an application of the method of iterative averages for quot; = 0:2, the distributions with the same total variance PK k=1 V (QCk; Pk): We can see di erent result in Table 12 where limit distributions of both modi cations were the same. This di erent behavior of the modi cations seems to be dependent on a shape of covering (C1; C2; : : : ; CK) of index set V which input distributions have.... ..."

### Table 4. Estimation of ED Cutoff Values from Linear Regression of Calculated and Empirical Molecular Volumes and Areas of Projectiona

2005

"... In PAGE 9: ... Based on this assumption, the linear regressions of calculated areas of projection (AED) and the areas calculated from an equation of state (AvdW ) 2/4) were performed (Table 4). Only a rough estimation of EDC may be obtained from the results in Table4 . Still, it confirms the previously used values at least to the order of magnitude and this is quite satisfactory considering the exponential character of the dependence of electronic densities on the distance from the centers of atoms.... ..."