### Table 1: Quadratic or Nonparametric?

2001

"... In PAGE 8: ... The squared L2 risks of the estimators are computed based on 100 replications. The numbers in the parentheses in Table1 are the corresponding standard errors. Quadratic regression works much better than the nonparametric alternatives for the rst two cases, but becomes much worse for the latter two due to lack of exibility.... ..."

Cited by 13

### Table 3 shows the errors of nonparametric regression. The

2000

"... In PAGE 6: ...Figure 5: Rule #28Finger 4#29 Figure 6: Rule #28Finger 5#29 Figure 7: Rule #28Finger 12#29 Figure 8: Rule #28Finger 15#29 Figure 9: Rule #28Finger 21#29 Figure 10: Rule #28Finger 25#29 Table3 : Error #28Shiritori#29 slice err. slice err.... ..."

Cited by 1

### Table A2: Nonparametric (Kernel) Regressions

### Table 4: Valuing American put options using nonparametric regression

"... In PAGE 10: ... For details about the smoothing spline, see the help file of Matlab. The results are in Table4 and Figure 2. Table 4: Valuing American put options using nonparametric regression ... ..."

### Table 1: Nonparametric Lag Selection for Lynx Data

2000

"... In PAGE 14: ... We follow the suggested procedure of the last section and use only the CAF P E1 and the CAF P E2a criteria and for reasons of comparison, the linear Schwarz criterion ARSC. Table1 summarizes the results for the lynx data. Except for the CAF P E1 criterion all criteria include lag 1 and 2 in their selection.... In PAGE 14: ... Recalling the results of the previous section, these lags for the CAF P E2a may be due to over tting. To decide whether the more parsimonious model is su cient, we investigated the residuals of all suggested models using the bandwidths of Table1 and conclude that lags 1 and 2 are su cient. A plot of the estimated regression function on a relevant grid is shown in Figure 5.... In PAGE 15: ...and 3 using AF P E1 while Yao and Tong (1994) found lags 1, 3 and 6 using cross-validation. Insert Table1 about here Applying our methods to daily exchange rate data poses a di erent challenge. While there are plenty of data (3212 observations), this bene t is compromised as the data is known to be highly dependent (although only weakly correlated) and therefore asymptotics kick in very slowly.... ..."

Cited by 7

### Table 10: Pinball loss comparison between the nonparametric quantile regression without (npqr) and with (npqrm) monotonicity constraints.

2006

"... In PAGE 27: ... Note that on the engines data set the monotonicity constraint is not perfectly satisfied. Table10 shows the average pinball loss comparison between the nonparametric quantile regression without (npqr) and with (npqrm) monotonicity constraints. See above for the notation of the table.... ..."

Cited by 7

### Table 4: Summary Comparison of Parametric versus Nonparametric Models

2003

"... In PAGE 19: ... How do the nonparametric estimates compare to the linear parametric estimates presented in Table 2 above? In order to facilitate a direct comparison of results, we present a summary of the goodness-of-flt14 of the parametric and nonparametric estimates (R2 and root mean square error (RMSE)), along with the calculated elasticities of crime with respect to alcohol availability computed at the mean number of licenses and crime rates, Ec;a in Table 4. An examination of Table4 reveals that the nonparametric model provides a better flt to the un- derlying relationship between crime rates and alcohol availability than the linear parametric model. Speciflcally, the R2s shown in Table 4 indicate that the nonparametric model explains at least twice as much of the variation in crime rates as the linear parametric model does.... In PAGE 19: ... An examination of Table 4 reveals that the nonparametric model provides a better flt to the un- derlying relationship between crime rates and alcohol availability than the linear parametric model. Speciflcally, the R2s shown in Table4 indicate that the nonparametric model explains at least twice as much of the variation in crime rates as the linear parametric model does. We also note that the regression standard errors are much lower for nonparametric estimates than their parametric counterparts.... ..."

### Table 3 In-sample analysis of mortgage price changes regressed on Treasury bond futures price changes: Rolling OLS and nonparametric regressions, 1982-1986.

### Table 1: Nonparametric lag selection for lynx data Est. method max. # lags Selected lags crit. value

2000

"... In PAGE 14: ... We follow the suggested procedure of the last section and use only the CAFPE1 and the CAFPE2a criteria and for reasons of comparison, the linear Schwarz criterion ARSC. Table1 summarizes the results for the lynx data. Except for the CAFPE1 criterion all criteria include lag 1 and 2 in their selection.... In PAGE 14: ...ags. Only the CAFPE2a additionally suggests lags 5 and 8. Recalling the results of the previous section, these lags for the CAFPE2a may be due to over tting. To decide whether the more parsimonious model is su cient, we investigated the residuals of all suggested models using the bandwidths of Table1 and conclude that lags 1 and 2 are su cient. A plot of the estimated regression function on a relevant grid is shown in Figure 5.... ..."

Cited by 7

### Table 3. Results of parametric and nonparametric tests for between-shape differences in fencerow and intersection characteristics.

"... In PAGE 9: ... Regression analyses indicated that none of the fencerow characteristics were related to the richness variables, implying that this control technique was effective and that we therefore obtained a clear assessment of the effect of inter- section shape on plant richness. Differences in intersection area among intersec- tion shapes ( Table3 ) were probably just a physical result of the number of intersecting fencerows comprising L, T and X intersections. One could argue, however, that vertebrate-dispersed richness may have been higher in intersections with more avenues for influx simply because such intersec- tions happened to be larger and therefore more structurally diverse (c$ Gutzwiller and Anderson 1987), which could have attracted a greater variety ... ..."