Abstract This paper examines the effect of intraday timing on market volatility in Indian equity markets, emphasizing the interaction between liquidity, information flow, and investor behavior. Using one and five-minute data from the National Stock Exchange (NSE) between 2018 and 2025, volatility is modeled through GARCH-type econometric models frameworks such as Wavelet Realized Volatility and LSTM-GARCH. The results reveal a distinct U-shaped intraday volatility curve with peaks at market opening and closing hours and heightened fluctuations during macroeconomic announcements. The hybrid LSTM-GARCH model demonstrates superior predictive accuracy, outperforming conventional GARCH by roughly 25 percent. Findings highlight that combining econometric structure with deep-learning flexibility improves real-time volatility forecasting in emerging markets like India.
Raj et al. (Sat,) studied this question.