
Research Article
Forecast of the VN30 Index by Day Using a Variable Dimension Reduction Method Based on Kernel Tricks
@INPROCEEDINGS{10.1007/978-3-030-92942-8_8, author={Thanh Do Van and Hai Nguyen Minh}, title={Forecast of the VN30 Index by Day Using a Variable Dimension Reduction Method Based on Kernel Tricks}, proceedings={Nature of Computation and Communication. 7th EAI International Conference, ICTCC 2021, Virtual Event, October 28--29, 2021, Proceedings}, proceedings_a={ICTCC}, year={2022}, month={1}, keywords={Time series Dimensionality reduction Kernel tricks PCA Stock index forecast}, doi={10.1007/978-3-030-92942-8_8} }
- Thanh Do Van
Hai Nguyen Minh
Year: 2022
Forecast of the VN30 Index by Day Using a Variable Dimension Reduction Method Based on Kernel Tricks
ICTCC
Springer
DOI: 10.1007/978-3-030-92942-8_8
Abstract
The stock market is an important channel in attracting investment capital. Its stock index and the stock prices of some large-cap enterprises are often leading indicators of the economy. Stock index forecasting is always a matter of interest to researchers and applications. There is now much research on stock market forecasting using technical or fundamental analysis methods through various soft computing techniques or algorithms.
In Vietnam, the VN30 stock index is calculated from the trading results of the 30 largest capitalization enterprises of the economy. The capitalization rate of these 30 enterprises accounts for more than 80% of the capitalization rate of the stock market. Investors and economic analysts often use this index's forecast information when making investment decisions or to analyze and forecast the activities of the economy. However, the information and data affecting the fluctuation of the VN30 stock index are often very large, which is the main obstacle when forecasting this index.
The purpose of this article is to perform a daily forecast of the VN30 index on a large dataset of predictors by using econometric techniques on factors extracted from the dataset of the predictors using the variable dimension reduction method based on kernel tricks. The forecast accuracy according to such an approach is very high.