Please use this identifier to cite or link to this item:
https://ptsldigital.ukm.my/jspui/handle/123456789/485838Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor.advisor | Abu Hassan Shaari Mohd Noor, Prof. Dr. | |
| dc.contributor.author | Ehsan Hosseinidoust | |
| dc.date.accessioned | 2023-10-10T09:06:26Z | - |
| dc.date.available | 2023-10-10T09:06:26Z | - |
| dc.date.issued | 03/07/2013 | |
| dc.identifier.other | ukmvital:77066 | |
| dc.identifier.uri | https://ptsldigital.ukm.my/jspui/handle/123456789/485838 | - |
| dc.description | Stock market forecasting still remains a challenging and complicated problem for individual investors as well as corporations and authorities. The main aim of forecasting is to decrease the risk associated with investment thereby contributing to investment utility maximization. For this purpose, different models and methods have been developed based on the concepts of time series modeling and Artificial Intelligence (AI) methods. However, there is still no consensus on the accuracy of such procedures in stock market forecasting. Therefore, this study intends to compare the prediction performance of alternative modeling approaches based on relevant theories such as the Efficient Market Hypothesis and the Arbitrage Pricing Theory in ASEAN-5 countries as well as investigate the informational efficiency of ASEAN-5 stock markets. Thus, this research covers the two popular viewpoints on stock market analysis, namely technical analysis and fundamental analysis, which are represented by univariate and multivariate modeling respectively. Results of this research point out several interesting features in ASEAN-5 stock markets; first, in the case of univariate time series modeling, nonlinear structures have higher precision than the linear models such as ARIMA, particularly in the out-sample forecasting. Secondly, results of forecasting using AI methods indicated higher accuracy of these methods in comparison with the time series models for both univariate and multivariate scenarios and in terms of in-sample and out-sample forecasting. Thirdly, in the multivariate scheme, the FF-BP neural network and hybrid system both outperformed the VECM procedure, and the hybrid system (a combination of data-driven and model-driven methods) produces higher forecasting accuracy than the purely data-driven (FF-BP) method. In addition, the results of this study indicated that among the time series models, the univariate approaches provided higher forecasting accuracy than the multivariate model of VECM, in both in-sample and out-sample forecasting. Furthermore, results of the hybrid system in the out-sample forecasting is more accurate than the NARX network, supporting the application of combined models as the optimum approach for forecasting the stock market. For majority of ASEAN-5 stock markets, comparison between results of forecasting in the market level, sector level and firm level indicated that forecasting in the market level had higher accuracy in terms of in-sample and out-sample forecasting error criteria. Lastly, since the sample duration also included the period of financial crisis experienced by South East Asian countries, the reiterating test of cointegration using an advanced version of Johansen proved that the long-run relationship between stock market composite index and selected macroeconomic factors in the case of each country remained valid. Findings of this study imply that nonlinear methods are more capable of capturing the data generating process of the stock market and the forecasting results of these methods are more accurate than linear procedures. In addition, the VECM procedure is suitable for finding interactions between the stock market index and macro variables, whereas intelligence methods in general and hybrid system in particular, are more appropriate for forecasting purposes due to their flexible structures.,PhD | |
| dc.language.iso | eng | |
| dc.publisher | UKM, Bangi | |
| dc.relation | Faculty of Economy and Management / Fakulti Ekonomi dan Pengurusan | |
| dc.rights | UKM | |
| dc.subject | Modeling and forecasting | |
| dc.subject | Stock markets | |
| dc.subject | Asean | |
| dc.subject | Stock price forecasting | |
| dc.title | Modeling and forecasting of stock markets using linear and nonlinear methods : new evidence from ASEAN-5 countries | |
| dc.type | Theses | |
| dc.format.pages | 375 | |
| dc.identifier.callno | HG4637.H647 2013 tesis | |
| dc.identifier.barcode | 003202 (2013) | |
| Appears in Collections: | Faculty of Economy and Management / Fakulti Ekonomi dan Pengurusan | |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| ukmvital_77066+SOURCE1+SOURCE1.0.PDF Restricted Access | 2.91 MB | Adobe PDF |  View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.