The VIX Index is an indicator of the market’s perception of risk and an accurate forecast of the movements in VIX can be very useful for investment risk management. The aim of this study is to predict the day-to-day movement of the VIX index for Indian stocks using various deep learning algorithms. Daily data of India VIX and Nifty 50 was taken from NSE, and daily data of CBOE VIX and S&P 500 was taken from Yahoo Finance using yfinance Python module from March 2009 to April 2021. The Deep Learning architectures studied are Long Short-Term Memory (LSTM ), Gated Recurrent Unit (GRU ), bidirectional LSTM, bidirectional GRU, simple Recurrent Neural Network (RNN) and Conv1D (one dimensional Convolutional Neural Network). All six architectures performed well and achieved a higher level of accuracy than in previous studies. Accuracy score ranges from 63% to 65%, area under the Receiver Operating Characteristics curve ranges from 64% to 66%, and area under the Precision- Recall curve ranges from 60% to 65%. There were only minor differences among the six networks tested, although Simple RNN and Bidirectional LSTM show slightly better overall accuracy. However, Conv1D may be the preferred choice since it achieves comparable accuracy with a more limited hardware and takes less time to train. The findings of the study are of great relevance for assessing the short-term risk as well as long term strategies for hedgers, risk averse investors, volatility traders, investors, and financial researchers. Specifically, it would be vital for the investors and traders who are interested in investing money in India’s stock markets: the NSE of India and the Bombay Stock Exchange.
