Nonparametric path function estimation of Fourier series at low oscillations for modelling timely paying credit

The development of the nonparametric path model is carried out with the assumption that each function formed has the same data pattern shape. However, in actual cases, several cases are often encountered where the data patterns formed are different from each of the calculated functions. This research aims to estimate the nonparametric path function of the Fourier series and to describe the lemma and theorem for the analysis of the nonparametric path of the Fourier series at low oscillation levels (K = 2,3,4,5). Primary data is obtained from customers at a Bank (Bank X) in Indonesia. The function estimation in nonparametric path analysis using the Fourier series approach is α(λ)=(n-1X'X + λD)-1-n-1 X' y . The best nonparametric path model that can describe the 5C variable on Time to Pay through Willingness to Pay is when the oscillation K = 4 with R2 is 78%. This study applies the Fourier series approach to path analysis in the banking sector.