### Table 5 Cross Nested Logit model parameter estimates

"... In PAGE 12: ... We then constrained several parameters to a fixed value, in order to obtain an estimable model. The model reported in Table5 contains 14 estimated parameters. The final loglikelihood is... ..."

### Table 3: Conditional Logit Model of Region Choice

2002

"... In PAGE 20: ...Table3 ). We then turn to a nested logit specification in which we first estimate the choice of region within a given nation and then estimate the choice of nation taking into account the attractiveness of its constituent regions (Table 4).... In PAGE 20: ... 3.1 Region Choice: Non-nested Logit Table3 provides results for six different conditional logit estimations of the location choice of Japanese affiliates among EU regions. Specification (1) shows coefficients for the standard set of production costs variables, excluding the difficult-to-interpret agglomeration effects.... In PAGE 25: ... 3.2 Region choices nested within nation choices The use of the country dummies in specifications (5) and (6) of Table3 helps to mitigate the problem associated with non-independent errors across regions belonging to the same nation. However, the country fixed effects do not resolve problems associated with cross-industry and... In PAGE 26: ...22 choices per Japanese investor. Table 4 re-estimates the same specifications as Table3 in a nested structure. The results are broadly similar.... ..."

Cited by 3

### 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 3 Leave-one-out classification performances (percentages) for the LDA, Logit and LS- SVM model using the full candidate input set. The corresponding p-values (per- centages) are denoted between parentheses. LDA LOGIT LS-SVM

2004

"... In PAGE 25: ... The generalization performance is assessed by means of the leave-one-out cross-validation error, which is a common measure in the bankruptcy prediction literature [22]. In Table3 , we have contrasted the PCC, PCCp, PCCn and AUROC performance of the LS-SVM (26) and the Bayesian LS-SVM decision rule (59) classifier with the performance of the... ..."

### TABLE Standard Logit Model

1998

Cited by 16

### Table 2: Estimates of Logit Models of Rivalry Terminationsa Logit Logit with Time

"... In PAGE 17: ... He also hypothesized that the relative military power of the rivals (BALANCE OF POWER), war between the rivals (WAR) or the polarity of the international system (BIPOLAR) would not in uence rivalry termination. Table2 presents our reanalysis of Bennett apos;s results on rivalry terminations. The rst two columns show our reanalysis of Bennett apos;s original BTSCS analysis, with results identical to his published results.... ..."

### Table 5: Multinomial Logit Model relating Firm and Industry Characteristics to the

"... In PAGE 22: ... If conflicts of interest and lender liability are important considerations, the effects we identified above should be most pronounced when there is a lending relationship and particularly if the banker represents the main bank lender to the firm. 37 In panel A of Table5 , the dependent variable is classified into three categories: no banker on the board (n=294), a banker on the board with no lending relationship (n=29), a banker on the board with some lending relationship (n=48). In panel B, the third category is split to distinguish between a banker on the board with a main bank lending relationship (n=14) and a banker on board with a minor lending relationship (n=34).... In PAGE 23: ... 21 specifications include one-digit SIC code industry indicators but their coefficients are not reported. Table5 shows that as the lending relationship becomes more important, the estimated coefficients for the linear and non-linear volatility measures increase in absolute value and remain statistically significant. In terms of the absolute value of the point estimates, the 38 coefficients for the other variables are greater for the main banker group than for bankers with either minor or no banking relationship.... In PAGE 43: ...Table5 (continued) Panel B: Four Categories In panel A, the dependent variable is classified into three categories: no banker on the board (n=294), a banker on the board with no lending relationship (n=29), a banker on the board with some lending relationship (n=48). In panel B, the third category is split to distinguish between a banker on the board with a main bank lending relationship (n=14) and a banker on board with a minor lending relationship (n=34).... ..."

### Table 4 Leave-one-out classification performances for the LDA, Logit and LS-SVM model using the optimized input sets. LDA LOGIT LS-SVM

2004

"... In PAGE 27: ... Notice the similarities between both curves during the input removal process. Table4 reports the performances of all classifiers using the optimally pruned set of inputs. Again it can be observed that the LS-SVM and LS- SVMBay classifiers yield very good performances when compared to the LDA and Logit classifiers.... ..."

### Table 3. Multinomial Logit Model

2000

"... In PAGE 13: ... The variables and random effects included in this model are the result of a classical selection procedure that estimated alternative MMNL models and used a likelihood ratio test to select from them. A likelihood ratio test at the five percent level shows that this model fits significantly better than a simple MNL model (given in Table3 ). The table gives estimates of the standard errors of the coefficients for 250 replications, and also for 50 replications.... In PAGE 13: ... In general, using the quot;Asymptotic quot; covariance formula with 250 replications results in a ten to twenty percent underestimate of standard errors of coefficients, compared to the quot;Robust quot; formula. Table3 , Model 1, is a simple MNL model L (i;x,G07) fitted to the data; these estimates are taken from C Brownstone amp; Train (1998).... In PAGE 14: ... All the factors identified in their search were picked up by our procedure, as were some additional candidates. Table 4 gives a MMNL model which includes the six random effects identified as possibly significant by the artificial variable test in Table3 , using T-statistics greater than one in magnitude as the selection criterion. The MMNL estimates show that there is significant mixing in each of these factors.... ..."

Cited by 37