### Table 4: Performance of the First-Order Log-Domain Filter

2003

"... In PAGE 64: ... Filters with any gain will cause problems in the circuit as the gain parameters are crucial to the Meddis model. Table4 shows the improvement of using cascode mirrors to increase the output resistance of current mirrors used in the filter. Little DC gain was measured across the frequency range.... ..."

Cited by 2

### Table 1: shows how the numerically computed width of the layer depends on as ! 0+. The theory predicts that the width is of order log( ) as ! 0+.

1997

"... In PAGE 5: ... The layer is de ned to start and end when the nite element solution is bounded away from these values by a small number - here we choose 0:03. Selected results for the PN diode problem are presented in Table1 , these clearly show that the desired order of is present in the computed widths. We also observe that the number of nodes needed to compute successively more severe layers does not blow up.... ..."

Cited by 3

### Table 3 CPT for HeartDisease, H1: no second-order interactions on the log-linear model.

"... In PAGE 4: ... For example, in Figure 4, Smoking and eXercise both af- fect your chances for Heart disease. We presume that Smok- ing increases your chances of Heart disease, and eXercise de- creases them, with the probabilities given in the first 3 rows of the CPT in Table3 . If we also presume that they do not interact according to the logit model, then the logit model fills in the final row as indicated in the shaded box.... In PAGE 11: ...11 Appendix Derivation of Table3 : logit FOM Table 3 shows non-interaction between S and X at H un- der the logit model. Given the first 3 rows of the CPT at H, we can derive the last row according to the FOM (Equa- tion 5).... In PAGE 11: ...11 Appendix Derivation of Table 3: logit FOM Table3 shows non-interaction between S and X at H un- der the logit model. Given the first 3 rows of the CPT at H, we can derive the last row according to the FOM (Equa- tion 5).... In PAGE 11: ... a BP log AIPrB4H1CYS0X0B5 PrB4H0CYS0X0B5 AJ (9) BP log AIBM3 BM7 AJ AP A00BM8473 b BP log AIPrB4H1CYS1X0B5 PrB4H0CYS1X0B5 AJ A0a (10) AP log AIBM2 BM8 AJ B70BM8473 AP A00BM5390 c BP log AIPrB4H1CYS0X1B5 PrB4H0CYS0X1B5 AJ A0a (11) AP log AIBM6 BM4 AJ B70BM8473 AP 1BM2528 B5 PrB4H1CYS1X1B5 BP expB4aB7bB7cB5 1B7expB4aB7bB7cB5 (12) AP 0BM8750 1BM8750 AP .4667 Derivation of noisy-OR values for the Smoking model Given the ordering in Table3 , we can read off that q0 BP BM8 (the largest q has to be q0), so qX BP BM7BPq0 BP BM875, qS BP BM4BPq0 BP BM5, and PrB4H0CYX1S1B5 BP q0qXqS BP BM35. Therefore, PrB4H1CYX1S1B5 BP q0qXqS BP BM65.... ..."

### Table 4 Relative errors for the singular case s(x) = log x. Here the error is of order O(hl log h), where l is shown.

1999

Cited by 5

### Table 3: Log likelihood value of the ordered logit model for cross-buying

"... In PAGE 6: ...20)). The LL values of the three models are displayed in Table3 . Again, the inclusion of SP variables in the joint model is not significant at the 5% level, while the inclusion of RP variables is significant even at the 0.... ..."

### Table 1 The canonical coe cients indicating the generic magnitude of various leading and subleading contributions up to third order. The big log L is calculated for = 25 mrad.

### Table 4. Log-likelihood results in order of best fit. Model State-dependent Age-dependent No. parameters Log-likelihood

2006

"... In PAGE 6: ... In fact, as mentioned before, models A, B and C are special cases of model D. Table4 summarizes the characteristics of the four models and the results for the log- likelihood values. It can immediately be seen that the most general model, D, has the highest log- Table 4.... ..."

Cited by 1

### Table 3 Average results of the log-linear model with interaction terms of order 2 on all the features for Experiment I

"... In PAGE 7: ... We repeat the procedure ten times with different random splits and compute the averages. We analyzed models of various orders of feature interaction; Table3 shows the results for an order 2 model on all features. The regression-based model utilizes information from both within and outside the corpus of biological articles (by comparing verb frequencies with the corresponding fre- quencies in the Wall Street Journal and medical corpus) and optimizes the weights for the various counts and test probabilities according to the training data.... In PAGE 8: ...Table3 ). This demon- strates that when the labels of the verbs become more accurate, the performance of AVAD improves in both precision and recall when using the proportions test.... ..."

### Table 3. Test Sequence Log-Odds Scores for VOGUE, HMMER and k-th Order HMMs

"... In PAGE 10: ...Score Comparison: We first compare VOGUE with k-order HMMs and HM- MER. Table3 shows the comparison on the 5 test sequences for family F1 when scored against the model for F1. For VOGUE we used minsup = 27(75%) and maxgap = 20.... In PAGE 10: ... The best score for each sequence is highlighted in bold. In Table3 , we find that k-th order HMMs were not able to model the training sequences well. All their scores are large negative values.... ..."

### Table 3. Test Sequence Log-Odds Scores for VOGUE, HMMER and k-th Order HMMs

"... In PAGE 10: ...Score Comparison: We first compare VOGUE with k-order HMMs and HM- MER. Table3 shows the comparison on the 5 test sequences for family F1 when scored against the model for F1. For VOGUE we used minsup = 27(75%) and maxgap = 20.... In PAGE 10: ... The best score for each sequence is highlighted in bold. Looking at the scores in Table3 , we find that in general k-th order HMMs were not able to model the training sequences well. All their scores are large negative values.... ..."