Suparti* | Alan Prahutama | Rukun Santoso
OPEN ACCESS
The regression model can be approximated by parametric and nonparametric methods. The parametric regression method generates an excellent regression model when the shape of the curve is known, whereas if the curve shape is random, it can be approximated by using a nonparametric regression method. There is a nonparametric regression method that has been developed such as spline truncated and local polynomials. Spline truncated is segmented pieces regression, while the local polynomial is polynomial models with kernel function as weighted. The most important thing in nonparametric regression modeling is the selection of smoothing parameters. One of the selected parameters of the method is the Generalized Cross Validation (GCV) method. The purpose of this study is to generate the model of Inflation’s sectors in Indonesia using Spline Truncated and local polynomial. These sectors include foodstuffs sector; food, beverages, cigarettes, and tobacco sector; housing, water, electricity, gas, and fuel sector; clothing sector; health sector; education and sports sector; as well as transportation, communication, and financial services group. The results indicated that by modeling the value of the inflation sectors in Indonesia using Spline truncated resulted in average R-square is 68.86% while for local polynomial modeling, the average R-square is 73.73%.
spline truncated, local polynomial, inflation sectors in Indonesia
[1] Delaigle A, Fan J, Carroll RJ. (2009). A design-adaptive local polynomial estimator for the errors in variable problem. Journal of the Amercan Association 104 (485): 348-359. http://dx.doi.org/10.1198/jasa.2009.0114
[2] Welsh AH, Yee TY. (2005). Local regression for vector responses. Journal of Statistical Planning and Inference 136(9): 3007-3031. http://dx.doi.org/10.1016/j.jspi.2004.01.024
[3] Xue L. (2010). Empirical likelihood local polynomial regression analysis of clustered data. Scandinavian Journal of Statistics: Theory and Applications 37(4): 644-663. https://doi.org/10.1111/j.1467-9469.2009.00677.x
[4] Dette H, Melas VB. (2010). A note on all-bias designs with applications in spline regression models. Journal of Statistical Planning and Inference 140(7): 2037-2045. http://dx.doi.org/10.1016/j.jspi.2010.01.047
[5] Huang JZ. (2003). Asymtotics for polynomial spline regression under weak conditions. Statistics & Probability Letter 65(3): 207-216. http://dx.doi.org/10.1016/j.spl.2003.09.003
[6] Bryson KMO, Ko M. (2004). Exploring the relationship between information technology investment and firm performance using regression spline analysis. Information and Management 42(1): 1-13. https://doi.org/10.1016/j.im.2003.09.002
[7] Takezawa K. (2006). Introduction to Nonparametric Regression. John Wiley & Sons, New Jersey.
[8] Fransisco MF, Opsomer FJ, Fernandez MV. (2004). Plug in bandwidth selector for local polynomial regression estimator with correlated error. Nonparametric Statistics 16(1-2): 127-151. https://doi.org/10.1080/10485250310001622848
[9] Xiao Z, Linton OB, Carroll RJ, Mammen E. (2003).
[10] More efficient local polynomial estimation innonparametric regression with autocorelated errors. Journal of the American Statistical Assosiation 98(464):980-992. https://doi.org/10.1198/016214503000000936