Uji Linearitas Deret Berkala (Time Series) Memakai Rentang Kuartil Sampel

Abstract

In statistics, bispectrum is a statistic applied for finding out non linear interaction in statistical data. The inconsistency of bispectrum of time series is used to determine the non lincarity of the time series in stochastic process. This thesis analysed the size and the power of Hinich Licarity bispectral test. The modified test also uses the sample interquartile range by measuring the equality of location parameters. This thesis illustrates the limitation of the sample interquartile range as a measure of the hypothesis deviation