Page 191 - IJOCTA-15-1

Key drivers of volatility in BIST100 firms using machine learning segmentation
Complement this by focusing on post-socialist jumps are the primary drivers of the positive rela-
countries, using the Feasible Generalized Least tionship between total volatility and illiquidity. 24
Squares model to reveal significant volatility de- This insight is crucial for understanding the mi-
terminants such as economic freedom, turnover crostructure of financial markets. As the field of
ratio, and corruption levels. 16 Similarly, examines volatility analysis has evolved, researchers have
African stock markets, finding that dividend pay- increasingly turned to advanced computational
out, leverage, firm size, and earnings per share are techniques to enhance the accuracy and robust-
critical factors. 17 These studies collectively under- ness of their models.
score the importance of contextual factors in un- The role of managerial overconfidence in the rela-
derstanding stock price volatility. tionship between R&D volatility and firm value,
Further, (2020) investigates Indonesian firms, as examined by, provides insights into the Metal
showing that trading volume and firm size sig- Goods Sector of Borsa Istanbul. The study finds
nificantly affect volatility, while inflation and ex- a short-term positive link between R&D volatility
change rates do not. 18 This highlights that dif- and firm value, amplified by overconfident man-
25
ferent markets may have unique volatility deter- agers who view R&D risks as opportunities.
minants. Add another layer by exploring how the However, the study’s small sample and limited
COVID-19 pandemic impacted stock price volatil- scope caution against broad generalization.
ity, identifying firm size, dividend payments, Similarly, highlights trading volume as a proxy
and trading volume as significant determinants. 19 for unobservable information flows, impacting
This indicates that external shocks, such as pan- volatility in small-cap stocks during recessions. 26
demics, can alter traditional volatility dynamics. Reveals sectoral disparities in response to market
Having explored the various determinants of stock shocks, with financial leasing sectors being more
price volatility, we now focus on the methodolo- vulnerable. 27 These findings deepen our under-
gies used to measure and analyze this crucial mar- standing of market volatility, providing tools for
ket phenomenon. strategic decision-making and connecting broader
financial systems through regime-based insights
Transitioning from determinants to methodolo- 28
by.
gies, various techniques have been developed to
Explore the impact of investor learning from
measure and analyze stock market volatility. In-
past price observations on stock market volatility,
troduces multifactor volatility models, identifying
showing that subjective beliefs close to rational
long-term, medium-term, and short-term factors
influencing stock volatility. 20 These models pro- expectations impart momentum and mean rever-
29
sion into price-dividend ratios. This indicates
vide a comprehensive framework for understand-
that learning dynamics play a significant role in
ing volatility dynamics over different time hori-
market behavior, influencing volatility trends.
zons, offering practical tools for investors and risk
Applying machine learning and advanced regres-
managers.
sion techniques has significantly enhanced volatil-
Take a different approach by integrating wavelet ity analysis. Explores machine learning as an al-
decomposition, statistical learning, and econo- ternative to maximum likelihood estimation for
metric methods to examine the associations and GARCH(1,1) model parameters, showing its po-
causal influences among volatility indicators. 21 tential for improved predictive analysis. 30 This
This integrated framework enhances predictive suggests that machine learning techniques can en-
models for stock market volatility, aiding in- hance volatility forecasting accuracy, offering new
vestors and policymakers in making informed de- tools for risk management.
cisions. Similarly, employs linear regression and Use multiple linear regression models to de-
GARCH models to forecast volatility for the Dow velop a new measure for stock market volatil-
Jones Industrial Average, demonstrating the re- ity that outperforms traditional implied volatility
liability of these models in predicting market measures. 31 This innovation provides more accu-
volatility. 22 rate predictions, benefiting investors and finan-
Extends this analysis to euro currency pairs and cial analysts in their decision-making processes.
bond markets using GARCH models, revealing Similarly, proposes a hybrid support vector re-
interconnected volatilities influenced by macroe- gression model with chaotic genetic algorithms for
conomic events. 23 Understanding these connec- volatility prediction, demonstrating superior per-
tions helps investors and policymakers mitigate formance over standard models. 32
risks associated with currency and bond market Further, improve the accuracy of volatility fore-
volatilities. Further decompose total volatility casts using support vector regression with a hy-
into jump and diffusive components, showing that brid genetic algorithm. 33 This model is valuable
185