PEMODELAN DATA SAHAM MENGGUNAKAN ANALISIS TIME SERIES DENGAN PENDEKATAN COPULA GAUSSIAN

One method of predicting stock prices is to use the time series analysis method. In this method, a linear prediction model is made to see patterns from historical stock price data to assess future prices. The stock data used in this study is the daily stock data of PT. Telkom and PT. Indosat in 2020-2021. Autoregressive (AR) model is a time series model that is often used with the assumption that its volatility does not change with time (Homoscedastic). After analyzing the AR Model(1) data for the stock data of PT. Telkom and PT. Indosat has a non-independent error, therefore the AR(1)-N.GARCH(1,1) time series model construction was carried out to model the error (ϵ_(i,t)). Furthermore, the error of the AR(1)-N.GARCH(1,1) model is independent of t, so it can be modeled using Copula. After the Copula model was applied to the data and obtained the value of the fit of the Gaussian Copula distribution error model. From the values generated from the Gaussian Copula C({ϵ_(i,t) }_(t=1)^T ),T=1,2,…, and approximates a uniform distribution. So the stock data of PT. Telkom and PT. It can be said that Indosat is not suitable to be modeled with the Gaussian Copula.