The purpose of this study is to obtain the best approach in longitudinal data using Path Nonparametric Truncated Spline with PLS and PWLS approach. The data used in this study is longitudinal generation data. There are three variables, five levels of autocorrelation, and the data does not meet the assumption of linearity. The degree of autocorrelation is divided into very low autocorrelation (|0.01-0.20|), low (|0.21-0.40|), moderate (|0.41-0.60|), high (|0.61-0.80|), and very high (|0.81-1.00|). The analysis used is path nonparametric truncated spline with PLS and PWLS approach. The result of this research is the estimation of nonparametric path function using PLS and PWLS approach on various interactions. Based on the Relative Efficiency (ER) value obtained, PWLS approach is a better approach than PLS approach at all levels of autocorrelation when the number of observations in the subject is five (T=5). On the number of observations in fifteen and twenty subjects (T=15, 20), PWLS approach was better than PLS approach when the autocorrelation of the data was more than 0.20. The originality of this research is the development of a nonparametric path with the PLS and PWLS approach that is able to obtain the best approach in limitations of autocorrelation between observations and in one subject are interconnected. There has been no previous research that has examined the best approach of a path nonparametric with PLS and PWLS approach.
Digital Object Identifier (DOI)
Ajeng Puspitasari, Ananda; Achmad Rinaldo Fernandes, Adji; Efendi, Achmad; Astutik, Suci; and Sumarminingsih, Eni
"Development of Nonparametric Truncated Spline at Various Levels of Autocorrelation of Longitudinal Generating Data,"
Journal of Statistics Applications & Probability: Vol. 12:
2, Article 35.
Available at: https://digitalcommons.aaru.edu.jo/jsap/vol12/iss2/35