Predicting VN - Index Value by KNN Algorithm of Machine Learning

Abstract

The world has entered the stage of rapid development of technology, especially the fourth industrial revolution with outstanding changes and developments in information technology. Artificial Intelligence (AI) is one of the most mentioned names in this period. AI is part of computer science, developing technology in the direction of automation, self-learning. As a result, it takes a solid knowledge to be able to operate any AI system. There have been many applications of artificial intelligence in the fields of science, technology and economics -finance. From previous decades, the application of algorithms to predict values and variables in economics has been implemented and improved over many different stages. This paper aims to predict VNINDEX value by the application of KNN algorithm machine learning to assess the changes of price indexes and stock variables in general and the VNIndex stock exchange in particular. Research result shows that the outcome of a buy -or -sell decision at the point of view. With a predict signal value of 1, investors should execute buy and sell orders that are advised when predict signal yields -1. Overall, the result indicates that 51% of the stock market price are correctly predicted by the KNN algorithm machine learning.

Predicting S&P 500 Index ETF (SPY) During COVID-19 via K-Nearest Neighbors (KNN) Algorithm

“…Subasi et al (2021) find that within seven classifiers, including Random Forest, Bagging, AdaBoost, Decision Trees, SVM, KNN and ANN, KNN has the second highest accuracy (54%) in prediction for the NASDAQ dataset and has the highest accuracy for NIKKEI (56%) and FTSE (54%). Toai et al (2021) show that 51% of the stock market movement direction is correctly predicted by using the KNN algorithm to predict VN INDEX value from Ho Chi Minh Stock Exchange between 2013 and 2019. Kumbure et al (2022) also highlight the success of using KNN for stock market prediction conducted by Cao et al (2019) and Zhang et al (2017).…”

“…Previous research has demonstrated that the KNN method can generate better prediction accuracy than the time series method in economics, such as the linear AR-GARCH model (Meade, 2002). It has also been shown that the method performs better than other machine learning approaches, such as Logistic Regression, Linear Regression, Lasso, Elastic Net and Decision Tree Regression (Subha and Nambi, 2012;Toai et al, 2021). Subasi et al (2021) find that within seven classifiers, including Random Forest, Bagging, AdaBoost, Decision Trees, SVM, KNN and ANN, KNN has the second highest accuracy (54%) in prediction for the NASDAQ dataset and has the highest accuracy for NIKKEI (56%) and FTSE (54%).…”