### Table 4 Progress in Research on Forecasting Research from 1960 to 1984

### Table 1. Forecasting Performance of Grey-based RNN for Travel Time Estimation

"... In PAGE 11: ...issing data treatment is 0.00133, while the RNN without missing data treatment got 0.00154 MSE. Table1 lists the performance results of the grey-based RNN for estimating travel times for various prediction horizons (i.... ..."

### TABLE III THE ARIMA FORECASTING MODELS

2006

Cited by 1

### Table 1. Candidate Technologies for Diagnosis and Forecast of Terminal Area Ceiling and Visibility

"... In PAGE 23: ... end, a number of parallel forecast research efforts have begun to investigate and address the various needs. A list of candidate technologies that have been under consideration is provided in Table1 . Technologies that rely on the availability of high-resolution (time and space) observations such as satellite imagery, radar data, and surface observations will be used to address the tactical forecast time frame (less than three hours), while the longer forecast horizons needed for strate- gic planning (out to twelve hours) will rely more on the holistic approach of numerical weather prediction models.... ..."

### Table 4: Applications of New Method to Data from Mortality, Morbidity and Service Utilization Survey, Haiti, 1987.

"... In PAGE 9: ... It is therefore possible to apply the technique proposed here to data from the 1987 survey, as well as applying the age-based indirect method, and to compare the resulting estimates to both direct and age-based indirect estimates from the 1994 survey. Table4 shows the application of the new method to data on proportions dead tabulated by time since first birth. Coefficients from Tables 2 and 3 for the quot;West quot; family of Coale-Demeny model life tables have been used, since the direct birth histories from the 1994 survey suggest a close fit of child mortality patterns to that family.... ..."

### Table 3: The evolved grey-box model parameters

"... In PAGE 6: ... The parameters can be evolved in floating-points via Monte Carlo perturbations for varying the search domains. The parameters of the grey-box model resulting from the evolution are shown in Table3 . The modelling errors at different operating points are depicted in Fig.... ..."

### Table 1: Comparisons of Forecasts

"... In PAGE 10: ... For Model 3 t is uniform on the interval described by inequality (16), (14) and (15) are matched with = 0:6 and r = 1;; moreover we performed simulation runs with =0:3, where (15) is not met with Model 3. The results of this analysis are presented in Table1 . To measure the di erence in the performance of the forecasts we derive the coe cient of determination R 2 for Models 1 to 3, which corresponds to the percentage of the volatility of the returns explained by the statistical model.... In PAGE 10: ... To measure the di erence in the performance of the forecasts we derive the coe cient of determination R 2 for Models 1 to 3, which corresponds to the percentage of the volatility of the returns explained by the statistical model. The terms in parentheses in Table1 refer to the corresponding standard deviation of R 2 . Let us start with a smoothing parameter of = 0:6, i.... ..."