Using Machine Learning to Predict Missing Values in the Egyptian Stock Exchange

Document Type : Original Article

Authors

EL Madina Higher institute of Administration and Technology, Giza, Egypt

Abstract

Financial markets are very rich in information and variables. In contrast to the efficient market hypothesis, much research has been conducted to predict asset prices with promising accuracy. However, ensuring good models requires extracting important information from the given data sets. This paper examines the main Egyptian stock exchange index (EGX 30, EGX 50, EGX 70, EGX 100) and constructs some alternative portfolios by identifying important linear combinations of the EGX components. This can be dealt with by the missing data prediction technique, which is performed after principal component analysis (PCA). The results show that the main components, The results of the cross-validation (CV) of PCR show that the most important results emerge by analyzing the trends in INDEXOPEN, INDEXHIGH, INDEXLOW, and INDEXCLOSE over time, one can gain insights into the overall performance of the index. The values of the correlation coefficient range from -1 to 1. 1 means a perfect positive relationship, -1 means a perfect negative relationship, and 0 means no linear relationship. There are very strong correlations (close to 1) between INDEXOPEN, INDEXHIGH, INDEXLOW and INDEXCLOSE. This indicates that the values of these indicators move together significantly

Keywords

Main Subjects

Files

Share

How to cite

Statistics