No abstract

No abstract

No abstract

No abstract

No abstract

Financial markets across the globe have seen rapid volatility and uncertainty owing to scary and disruptive impacts of COVID-19 pandemic. Mayhem wrecked by frequent lockdowns, curfews, emergencies, etc. has stoked the high quantum of chaotic movement in equity markets and resulted in perplexed investor behaviour. It, therefore, is of paramount practical relevance to measure predictability of market fear at such a crucial juncture of time. Market fear can effectively be measured in terms of implied and historic volatility of equity markets. The present study chooses India VIX and 20-day rolling standard deviation of NIFTY returns to account for implied and historic volatility respectively during the ongoing COVID-19 timeline. Pertinent macroeconomic constructs, technical indicators and Google search volume index on meaningful keywords have been chosen as raw explanatory features for inspecting predictability. Boruta feature selection methodology has been used in a supervised manner to select significant features. State-Of-The-Art machine and deep learning algorithms namely Gradient Boosting (GB), Extra Tree Regression (ERT), Deep Neural Network (DNN), Long Short Term Memory Network (LSTM) are then used on processed feature set to scrupulously evaluate the quantum of predictability of said assets. The integrated predictive frameworks have been subjected to a battery of numerical and statistical checks to draw inferences. Additionally, Explainable AI frameworks have been used to analyse the nature of influence of respective features. Findings indeed suggest that despite exhibiting high degree of volatile traits, both India VIX and historic volatility can be predicted utilizing the proposed architectures effectively and serve practical actionable insights.

No abstract

No abstract