Non-transformed Principal Component Technique on Weekly Construction Stock Market Price

Authors

  • Yusrina Andu Department of Mathematical Sciences, Universiti Teknologi Malaysia, 81310 Johor Bharu, Johor, MALAYSIA
  • Muhammad Hisyam Lee Department of Mathematical Sciences, Universiti Teknologi Malaysia, 81310 Johor Bharu, Johor, MALAYSIA
  • Zakariya Yahya Algamal Statistics and Informatics Department, University of Mosul, Mosul, IRAQ

DOI:

https://doi.org/10.11113/matematika.v35.n2.1112

Abstract

The fast-growing urbanization has contributed to the construction sector becoming one of the major sectors traded in the world stock market. In general, non-stationarity is highly related to most of the stock market price pattern. Even though stationarity transformation is a common approach, yet this may prompt to originality loss of the data. Hence, the non-transformation technique using a generalized dynamic principal component (GDPC) were considered for this study. Comparison of GDPC was performed with two transformed principal component techniques. This is pertinent as to observe a larger perspective of both techniques. Thus, the latest weekly two-years observations of nine constructions stock market price from seven different countries were applied. The data was tested for stationarity before performing the analysis. As a result, the mean squared error in the non-transformed technique shows eight lowest values. Similarly, eight construction stock market prices had the highest percentage of explained variance. In conclusion, a non-transformed technique can also present a better result
outcome without the stationarity transformation.

Downloads

Published

2019-07-31

Issue

Section

Articles