BOHR International Journal of Finance and Market Research

2022, Vol. 1, No. 1, pp. 69–71

DOI: 10.54646/bijfmr.2022.11
www.bohrpub.com

METHODS

Predicting option prices and volatility with high-frequency

data using neural network
Weige Huang 1 and Hua Wang 2*
1 Wenlan School of Business, Zhongnan University of Economics and Law, Wuhan, China
2 School of Business, Shenzhen Technology University, Shenzhen, China

*Correspondence:
Hua Wang,
[email protected]

Received: 15 June 2022; Accepted: 30 June 2022; Published: 13 July 2022

A neural network utilizes a huge amount of data for analysis and prediction. This paper predicts option prices
and volatility using a neutral network based on high-frequency intraday data. We focus on short-term predictions
because option prices and volatility, in fact, are very volatile and almost impossible to predict. We find that neural
network can predict option prices and volatility by using predictors constructed from the prices of option and its
underlining index, especially in the short term, which is what practitioners care about more in practice.
Keywords: option, neural network, volatility, high-frequency data, price prediction

1. Introduction and literature review are available: weak hybrid models and strong hybrid models
(13, 14). The prediction based on the artificial neural network
An neural network mimics the brain neural network. The model for price movement is refined using traditional
artificial neural network contains an input layer, an output statistical models (14).
layer, and the inside layers. The network collects the There are various research works on the prediction of
information from the input layer and outputs the processed major stock indices. For instance, Yao et al. (15) predicted
information to provide useful outcomes. the price performance movements of the Nikkei 225 index
The artificial neural network model provides a new tool using an artificial neural network model. Gradojevic et al.
to study the price movement of financial products and (16) used an artificial neural network model based on the
economic indices, and numerous studies using the artificial data of expire time and the moneyless of the underlying
neural network models have been done on it (1, 2). Some instruments to predict the S&P-500 European options, and
studies use an artificial neural network model to focus on empirical tests provided the proper results for the option
GDP growth rate predictions, CPI rate predictions, and pricing. An artificial neural network model with multiple
other major economic indices (3–7). For economic activities, levels of functions is used by Morelli et al. (17) to predict
artificial neural network models have been used to predict the option price and forecast the hidden pricing movement
default and bankruptcy for consumer borrowings (8, 9). relationship between financial derivatives and the option-
For financial instrument price prediction, many studies related variables. Shakya et al. (18) used artificial neural
utilize a huge amount of financial data and try to forecast network models based on evolving algorithms to model
hidden relationships. For example, the price prediction of demand scenarios, and the results of the approach provided
the financial derivative using an artificial neural network the proper price accuracy of the price predictions compared
model can provide suitable results compared with a closed- with the traditional closed-formed models.
formed option model (10). Various researchers are using There are numerous studies on option pricing based on
artificial neural network models for option pricing (11, literature reviews, and our approach in this study applied
12). Considering the degree of using the integration of the an artificial neural network model for the option pricing
artificial neural network model, two major types of models movements and compared it with the short-term and

69
70 Huang and Wang

FIGURE 1 | True and predicted close prices (in sample). FIGURE 3 | True and predicted volatility (in sample).

FIGURE 2 | True and predicted close prices (out of sample).

FIGURE 4 | True and predicted volatility (out of sample).

long-term periods. This paper focuses on the short-term

prediction because option prices and volatility, in fact, are This paper uses the index price, index returns, rolling
very volatile and almost impossible to predict. The results volatility of index returns, implied volatility, and implied
show that neural network can predict option prices and
prices of the option to predict option prices by the neural
volatility by using predictors constructed from the prices
network. We use the first half of the dataset to train the
of option and its underlining index, especially in the short
model and the second half to test the model. We find that
term, which is what financial practitioners care about more
the neural network is expected to fit the data quite well in
in financial market trading.
the training set, as shown in Figure 1. Figure 2 shows that
the model based on our predictors can predict the option
prices in the short term but cannot predict the prices well in
2. Data and results the long term. Figures 3, 4 show the in-sample and out-of-
sample predictions of neural networks on option volatility.
This paper studies the predictability of option close prices
We can see that the predicted volatility fits the true volatility
by using the “IO2003C4000” call option prices. That being
quite well, and the predicted one is able to predict the trend
said, the 2003 series of call option’s underlined index is the
of the option volatility out of sample.
CSI 300 Index and the strike price is 4000. The expiration
We also conduct research on predicting option returns.
date of the option is March 20, 2020, the third Friday of
We find that a neural network with predictors constructed
the call option contract’s expiration month. We also use
the CSI 300 Index prices and other variables to construct as above is not able to predict the option returns. We
several predictors that are used to predict the call option conjecture that is because the option return is very volatile
volatility by a neural network. Our dataset’s frequency is and computing return by interdifferentiation makes the
half an hour (30 min), ranging from December 23, 2019, process lose a lot of memory, which is very important for
to March 20, 2020. In particular, we compute the following the prediction. We are able to predict prices because the
predictors: returns on the index, rolling volatility of index price process keeps the memory. One might suspect that the
returns, implied volatility, and implied prices of the option. price process is unstable in the long run. We note that we
Note that we set the risk-free rate to be 2% per year, and focus on short-run prediction in the option cases. Therefore,
assume no dividend when computing these predictors. The the unstable price process is not a big issue in this case. In
strike price is 4000. We compute the rolling volatility of the addition, the results on volatility prediction show that neural
index returns using the prior 24 observations. network is able to predict the uncertainty in option markets.
10.54646/bijfmr.2022.11 71

2. Shachmurove Y. Business applications of emulative neural networks. Int

3. Conclusion J Bus. (2005) 10:303–22.
3. Plasmans J, Verkooijen W, Daniels H. Estimating structural exchange
The artificial neural network model provides a new tool rate models by artificial neural networks. Appl Finan Econ. (1998)
to study the price movement of financial products. Our 541–51. doi: 10.1080/096031098332844
approach in this study applied an artificial neural network 4. Qi M, Wu Y. Nonlinear prediction of exchange rates with monetary
model for the option pricing movements and compared it fundamentals. J Emp Fin. (2003) 10:623–40. doi: 10.1016/S0927-
5398(03)00008-2
with the short-term and long-term time periods. This paper
5. Saltoglu B. Comparing forecasting ability of parametric
focuses on the short-term prediction because option prices and non-parametric methods: an application with Canadian
and volatility, in fact, are very volatile and almost impossible monthly interest rates. Appl Finan Econ. (2003) 13:169–76.
to predict. The results show that a neural network can predict doi: 10.1080/09603100110111259
option prices and volatility by using predictors constructed 6. Binner JM, Bissoondeeal RK, Elger T, Gazely AM, Mullineux AW.
from the prices of option and its underlining index, especially A comparison of linear forecasting models and neural networks: an
application to Euro inflation and Euro divisia. Appl Econ. (2005) 37:665–
in the short term. 80. doi: 10.1080/0003684052000343679
Since the Shanghai Shenzhen 300 stock index option 7. Dutta S, Kamstra M. Bond-rating: a non-conservative application of
started only in December 2019 and not much study has been neural networks. Proceedings of the IEEE International Conference on
done on the artificial neural network model for forecasting Neural Networks. San Diego, CA: (1988). p. 443–50. doi: 10.1109/ICNN.
the stock index call option prices, our research approach 1988.23958
develops an approach to examine the Chinese market, which 8. Ahn BS, Cho SS, Kim CY. The integrated methodology of rough set
theory and artificial neural network for business failure prediction.
is an imperfect market. The test result of our approach Expert Syst With Appl. (2000) 18:65–74. doi: 10.1016/S0957-4174(99)
indicates that the neural network with predictors constructed 00053-6
above is not able to predict the option returns in the 9. Altman E, Marco G, Varetto F. Corporate distress diagnosis:
stock index option, and we conjecture that is because the Comparisons using linear discriminant analysis and neural networks. J
Bank Finan. (1994) 18:505–29. doi: 10.1016/0378-4266(94)90007-8
option return is very volatile and computing return by
10. Tsenga C-H, Chenga S-T, Wang Y-H, Peng J-T. Artificial neural network
interdifferentiation makes the process lose a lot of memory,
model of the hybrid EGARCH volatility of the Taiwan stock index option
which is very important for the prediction. In contrast, we prices. Physica A. (2008) 387:3192–200. doi: 10.1016/j.physa.2008.
are able to predict prices because the price process keeps the 01.074
memory. One might suspect that the price process is unstable 11. Hutchinson J, Lo A, Poggio T. “A nonparametric approach to pricing
in the long run. We note that we focus on the short-run and hedging derivative securities via learning networks. J Finan. (1994)
49:851–89. doi: 10.1111/j.1540-6261.1994.tb00081.x
prediction in the option cases. Therefore, the unstable price
12. Malliaris M, Salchenberger L. A neural network model for estimating
process is not a big issue in this case. In addition, our results
option prices. Appl Intell. (1993) 3:193–206. doi: 10.1007/BF00871937
show that a neural network can predict option volatility. 13. Andreou P, Charalambous C, Martzoukos S. Pricing and trading
For future improvement of our artificial neural network european options by combining artificial neural networks and
model, we will combine the artificial neural network with parametric models with implied parameters. Eur J Oper Res. (2008)
traditional statistical models, such as GARCH for the 185:1415–33. doi: 10.1016/j.ejor.2005.03.081
optimization of the stock index options, and we will examine 14. Wang YH, Lin CT. The political uncertainty and stock market behavior
in emerging democracy: the case of Taiwan. Qual Quant. (2009) 43:237–
the data various financial markets and further study the 48. doi: 10.1007/s11135-007-9102-6
long-term trend in terms of the price movements in the 15. Yao J, Li Y, Tan C. Option price forecasting using neural networks.
derivatives markets. Omega. (2000) 28:455–66. doi: 10.1016/S0305-0483(99)00066-3
16. Gradojevic N, Gencay R, Kukolj D. Option pricing with modular neural
networks. IEEE Trans Neural Netw. (2009) 20:626–37. doi: 10.1109/
TNN.2008.2011130
Conflict of interest 17. Morelli M, Montagna G, Nicrosini O, Treccani M, Farina M, Amato P.
Pricing financial derivatives with neural networks. Physica A Stat Mech
The authors declare that the research was conducted in the Appl. (2004) 338:160–5. doi: 10.1016/j.physa.2004.02.038
absence of any commercial or financial relationships that 18. Shakya S, Kern M, Owusu G, Chin C. Neural network demand models
could be construed as a potential conflict of interest. and evolutionary optimisers for dynamic pricing. Knowledge Based Syst.
(2012) 29:44–53. doi: 10.1016/j.knosys.2011.06.023
19. Berestycki H, Busca J, Florent I. Computing the implied volatility in
stochastic volatility models. Commun Pure Appl Math. (2004) 57:1352–
References 73. doi: 10.1002/cpa.20039
20. Huh J. Pricing options with exponential Lévy neural network.
1. Medeiros MC, Teräsvirta T, Rech G. Building neural network models for Expert Syst With Appl. (2019) 127:128–40. doi: 10.1016/j.eswa.2019.
time series: a statistical approach. J Econ. (2005) 25:49–75. 03.008