### Table A4.2. Sectoral Shares for Simulations, 2001-2015 Final share of total in 2015

2004

### Table 3: The experimental results on full datasets and reduced datasets obtained by VS-SSVM. The 1st VS-SSVM means to run VS-SSVM once on the full data.

2003

"... In PAGE 10: ... Column 5 (Table 2) gives the results obtained by iteratively applying VS-SSVM to the original or reduced data until no variables have weights that flip signs. Table3 gives a comparison of the results based on all attributes versus those selected by one iteration of VS-SSVM. We compared VS-SSVM to the correlation coefficient ranking method.... In PAGE 10: ... The standard deviation of Q2, Std(Q2), is computed as the standard deviation of the squared errors on the test data scaled by the variance of the actual response. By cross referencing with Table3 , Table 4 shows that the ranking scheme by correlation coefficients failed to improve the generalization performance. The significance of the differences was assessed using a paired t-test.... ..."

Cited by 28

### Table 3: The experimental results on full datasets and reduced datasets obtained by VS-SSVM. The 1st VS-SSVM means to run VS-SSVM once on the full data.

2003

"... In PAGE 10: ... Column 5 (Table 2) gives the results obtained by iteratively applying VS-SSVM to the original or reduced data until no variables have weights that flip signs. Table3 gives a comparison of the results based on all attributes versus those selected by one iteration of VS-SSVM. We compared VS-SSVM to the correlation coefficient ranking method.... In PAGE 10: ... The standard deviation of Q2,Std(Q2), is computed as the standard deviation of the squared errors on the test data scaled by the variance of the actual response. By cross referencing with Table3 , Table 4 shows that the ranking scheme by correlation coefficients failed to improve the generalization performance. The significance of the differences was assessed using a paired t-test.... ..."

Cited by 28

### Table 3: The experimental results on full datasets and reduced datasets obtained by VS-SSVM. The 1st VS-SSVM means to run VS-SSVM once on the full data.

2003

"... In PAGE 10: ... Column 5 (Table 2) gives the results obtained by iteratively applying VS-SSVM to the original or reduced data until no variables have weights that flip signs. Table3 gives a comparison of the results based on all attributes versus those selected by one iteration of VS-SSVM. We compared VS-SSVM to the correlation coefficient ranking method.... In PAGE 10: ... The standard deviation of Q2,Std(Q2), is computed as the standard deviation of the squared errors on the test data scaled by the variance of the actual response. By cross referencing with Table3 , Table 4 shows that the ranking scheme by correlation coefficients failed to improve the generalization performance. The significance of the differences was assessed using a paired t-test.... ..."

Cited by 28

### Table 7-4. Worldwide Cargo RTM and ATM Estimates for 1995 to 2015

"... In PAGE 12: ...able 6-8. DOC/ATM at the International Carrier Level........................................................... 6-7 Table7 -1.... In PAGE 12: ...able 7-1. Demand Variables..................................................................................................... 7-2 Table7 -2.... In PAGE 12: ...able 7-2. Baseline Assumptions............................................................................................... 7-3 Table7 -3.... In PAGE 12: ...able 7-3. U.S. Cargo RTM and ATM Estimates for 1995 to 2015.......................................... 7-4 Table7 -4.... In PAGE 78: ....S. regions, respectively. The econometric results are presented in Table 7-1. Table7 -1. Demand Variables Region Variable Name Coefficient T-ratio World Real income LNGDP 2.... In PAGE 79: ... These assumptions are summarized in Table 7-2. Table7 -2. Baseline Assumptions Region Variable Annual growth rate World Real income 3.... In PAGE 79: ...30% Real yield -1.00% Starting with 1995 traffic totals, and applying the assumptions of Table7 -2 to the coefficients of Table 7-1, we generated forecasts of cargo traffic for each region for the 20-year period 1996 through 2015. We project that world cargo traffic will grow at an annual rate of 7.... In PAGE 79: ...30% Real yield -1.00% Starting with 1995 traffic totals, and applying the assumptions of Table 7-2 to the coefficients of Table7 -1, we generated forecasts of cargo traffic for each region for the 20-year period 1996 through 2015. We project that world cargo traffic will grow at an annual rate of 7.... In PAGE 80: ... The annual data is shown are Table 7-3. Table7 -3.... ..."

### Table 7-4. Worldwide Cargo RTM and ATM Estimates for 1995 to 2015

1998

"... In PAGE 12: ...able 6-8. DOC/ATM at the International Carrier Level........................................................... 6-7 Table7 -1.... In PAGE 12: ...able 7-1. Demand Variables..................................................................................................... 7-2 Table7 -2.... In PAGE 12: ...able 7-2. Baseline Assumptions............................................................................................... 7-3 Table7 -3.... In PAGE 12: ...able 7-3. U.S. Cargo RTM and ATM Estimates for 1995 to 2015.......................................... 7-4 Table7 -4.... In PAGE 78: ....S. regions, respectively. The econometric results are presented in Table 7-1. Table7 -1. Demand Variables Region Variable Name Coefficient T-ratio World Real income LNGDP 2.... In PAGE 79: ... These assumptions are summarized in Table 7-2. Table7 -2. Baseline Assumptions Region Variable Annual growth rate World Real income 3.... In PAGE 79: ...30% Real yield -1.00% Starting with 1995 traffic totals, and applying the assumptions of Table7 -2 to the coefficients of Table 7-1, we generated forecasts of cargo traffic for each region for the 20-year period 1996 through 2015. We project that world cargo traffic will grow at an annual rate of 7.... In PAGE 79: ...30% Real yield -1.00% Starting with 1995 traffic totals, and applying the assumptions of Table 7-2 to the coefficients of Table7 -1, we generated forecasts of cargo traffic for each region for the 20-year period 1996 through 2015. We project that world cargo traffic will grow at an annual rate of 7.... In PAGE 80: ... The annual data is shown are Table 7-3. Table7 -3.... ..."

### Table 3: The experimental results on full datasets and reduced datasets obtained by VS- SSVM. The 1st VS-SSVM means to run VS-SSVM once on the full data.

2003

"... In PAGE 11: ... The resulting numbers of variables are in the 3rd column of Table 2. VS-SSVM greatly reduced the number of attributes as shown in columns 4 and 5 of Table 2 while improving generalization (see Table3 ). Column 5 (Table 2) gives the results obtained by it- eratively applying VS-SSVM to the original or reduced data until no variables have weights that flip signs.... In PAGE 11: ...351 0.046 Table3 gives a comparison of the results based on all attributes versus those selected by one iteration of VS-SSVM. We compared VS-SSVM to the correlation coefficient ranking method.... In PAGE 12: ... The standard deviation of Q2, Std(Q2), is computed as the standard deviation of the squared errors on the test data scaled by the variance of the actual response. By cross referencing with Table3 , Table 4 shows that the ranking scheme by correlation coefficients failed to improve the generalization performance. The significance of the differences was assessed using a paired t-test.... ..."

Cited by 28

### Table 7-3. U.S. Cargo RTM and ATM Estimates for 1995 to 2015

"... In PAGE 12: ...able 6-8. DOC/ATM at the International Carrier Level........................................................... 6-7 Table7 -1.... In PAGE 12: ...able 7-1. Demand Variables..................................................................................................... 7-2 Table7 -2.... In PAGE 12: ...able 7-2. Baseline Assumptions............................................................................................... 7-3 Table7 -3.... In PAGE 12: ...able 7-3. U.S. Cargo RTM and ATM Estimates for 1995 to 2015.......................................... 7-4 Table7 -4.... In PAGE 78: ....S. regions, respectively. The econometric results are presented in Table 7-1. Table7 -1. Demand Variables Region Variable Name Coefficient T-ratio World Real income LNGDP 2.... In PAGE 79: ... These assumptions are summarized in Table 7-2. Table7 -2. Baseline Assumptions Region Variable Annual growth rate World Real income 3.... In PAGE 79: ...30% Real yield -1.00% Starting with 1995 traffic totals, and applying the assumptions of Table7 -2 to the coefficients of Table 7-1, we generated forecasts of cargo traffic for each region for the 20-year period 1996 through 2015. We project that world cargo traffic will grow at an annual rate of 7.... In PAGE 79: ...30% Real yield -1.00% Starting with 1995 traffic totals, and applying the assumptions of Table 7-2 to the coefficients of Table7 -1, we generated forecasts of cargo traffic for each region for the 20-year period 1996 through 2015. We project that world cargo traffic will grow at an annual rate of 7.... In PAGE 81: ...7-5 Table7 -4. Worldwide Cargo RTM and ATM Estimates for 1995 to 2015 Year Worldwide RTMs (billions) Worldwide ATMs (billions) DOC (cents/ATM) 1995 64.... ..."