ResearchBib Share Your Research, Maximize Your Social Impacts
Sign for Notice Everyday Sign up >> Login

Subsidizing Black Scholes Inefficiency Using ANN in Nifty Index

Journal: International Journal of Science and Research (IJSR) (Vol.9, No. 10)

Publication Date:

Authors : ;

Page : 1472-1476

Keywords : volatility; Black-Scholes; GARCH; ANN; error metrics; NIFTY index options;

Source : Downloadexternal Find it from : Google Scholarexternal


Purpose: The aim of this paper is to build a model using the artificial neural network (ANN) to eliminate or subsidize the incongruence in Black Scholes fair value and actual market value. Also, this paper aims to provide a review of the diffuse and scattered literature on this particular theme. Research gap- Although many linear and non- linear, stochastic models including jump and diffusion has been used to capture volatility and measure the efficiency of Black-Scholes Model (BSM), but none of them have been substantially correct and consistent. This study uses ANN to reduce the inaccuracy of the given model. Research methodology- To examine the aforesaid purpose CNX NIFTY 50 call and put options data have been collected over a period of 6 months beginning from 1st Jan 2019 to 30th June 2019. Also, for capturing volatility closing price of the index is collected for a period of 1 year (1st July 2018 to 30th June 2019). GARCH volatility is computed to bring out the volatility parameter of the model. Later, a feed-forward algorithm/ back propagation algorithm will be used to process the ANN model. For testing the accuracy of the models error metrics will be computed (mean error, the total mean squared error. Root mean squared error). Excel and R are the used workspace. Findings- The result shows that ANN model outperforms the classic Black Scholes Model. For comparing the performance of the two model, error metrics (MSE, RMSE) are used which reflects that ANN reports minimum error. Originality- Neural network in options pricing in India is at blooming phase. This paper will be an addition in this genre. These neural network techniques do not presumes any relationship among the variables, but most of the work in this area either follows linear or nonlinear relationship among the variables. Practical implications- This paper will try to educate the options trader about the uses of neural networks to reduce the systematic biases in the most celebrated option

Last modified: 2021-06-28 17:13:38