Stochastic modelling of interest rates of interbank lending in Latvia using local regression method

Abstract

The interest rates of the Latvian interbank lending market are consideved using the method of locally weighted regression of the first order. A model for interest rates is proposed and compared in prognosis quality with two standard lineas models: random walk and autoregressin. The main result of the work is the doubtless advantage of the nonparameter model over other competitive models, which indicates the presence of nonlinearity in the dynamics of the Latvian interest rates.