Research on predicting future market prices using similarity

Abstract

Many methods have been proposed to forecast future exchange rates. Among them, one of the methods is to use the data interval similar to the most recent data extracted from the past data for forecasting. In this study, we measure the accuracy of forecasting data based on the assumption that the data of past exchange rate fluctuations that are similar to the most recent data are extracted and that the future fluctuations will be the same as the most recent data. The number of data used and the period of forecasting are varied, and each future forecast is compared with the actual price, aiming at a higher forecasting accuracy rate.