Stock trading is a highly demanding task consisting of many different and connected parts, and that has up to recently been a domain reserved for professional stockbrokers (Abdolmohammadi & Sultan, 2002). Additionally, the application of machine learning and algorithmic trading systems has further changed the way stock markets operate with algorithms now generating the most of the trading volumes in equity futures.
While algorithmic trading has brought in benefits like reduced cost and reduced latency, it also brings with it enormous challenges for individual and retail investors who do not have the necessary technology to build such systems (Shah, Isah, & Zulkernine 2019).
Analyzing price behaviours and stock market movements is extremely challenging and the powerful algorithms which have been used for trading in the markets to generate high profits are kept confidential and proprietary and the research or methodology behind them is mostly never published (Shah, Isah, & Zulkernine 2019).
While we may not manage to completely fill this void, it is our desire to develop a model that can be easily used by individual and retail investors and that will use various machine learning and deep learning techniques to predict the movement of stock prices and to determine whether to trade or not.
Many widely recognized empirical studies show that financial markets are predictable to some extent (Chong et al., 2017 as cited in Shah, Isah, & Zulkernine 2019). In this section, we review the studies previously done in this area and also gather information about the mechanisms that are currently being used for the prediction of Stock market (Sadia et al., 2019).
Myer (2011) observed that analysts utilize various techniques in trying to determine where the stock price might be headed. Some of these techniques include time series forecasting, technical analysis, and machine learning modelling (Sadia et al., 2019).
Shen et al. (2012) write that the potential in forecasting financial markets by Machine learning algorithms have been widely studied. These algorithms try to forecast the movement of stock prices based on training given with the past value movements (Prasanna & Ezhilmaran, 2016).
Majaski (2019) notes that Moving averages are indicators that analysts can use to assess the trend of the stock by averaging the daily price over a fixed period. They achieve this by generating Buy and sell signals when a moving average of a shorter duration crosses that of a longer duration.
Time Series analysis considers time to be a very important parameter to generate series of stock price movement (Prasanna & Ezhilmaran, 2016). Ariyo et al.  recommend the use of autoregressive integrated moving average model (ARIMA) for predicting stock prices.
Katariya & Jain (2019) define the Root Mean Squared Error (RMSE) as the square root of the average of the square of the total error. They write that the use of RMSE is prevalent and that it makes an excellent general purpose error metric for numerical predictions.
Analysing stock market movements and price behaviours is extremely challenging because of the markets noisy, dynamic, nonstationary, nonlinear, nonparametric, and chaotic nature (Abu-Mostafa and Atiya 1996). According to Zhong and Enke (2017), stock markets are affected by many highly interrelated factors that include company-specific variables, political, economic, and psychological (Park and Irwin 2007; Nguyen et al. 2015). [(Shah, Isah, & Zulkernine 2019).]
If you need assistance with writing your essay, our professional essay writing service is here to help!Essay Writing Service
There are two main approaches to analyse the financial market namely Fundamental analysis and Technical analysis (Majaski, 2019; Irwin 2007; Nguyen et al. 2015). [(Shah, Isah, & Zulkernine 2019).] Fundamental analysis forecast by attempting to measure the intrinsic value of the stock, while Technical analysts identify trends and patterns that suggest what direction a stock will take in the future (Majaski, 2019).
Fundamental analysts investigate many things ranging from the overall industry and economy conditions to the management and financial condition of companies. Assets, liabilities, earnings, and expenses, are all important characteristics to fundamental analysts. Technical analysis on the other hand only uses the stock's price and volume as the only inputs with the assumption that all known fundamentals are factored into the price (Majaski, 2019).
Sadia et al., (2019) writes that there are various methods and ways of implementing a prediction system. Due to the vast number of options available, there can be lots of models that can be used to predict the price of the stock. Furthermore, More and more researchers are investing their time every day in coming up with ways to arrive at techniques that can further improve the accuracy of stock prediction models (Sadia et al., 2019).
Whatever the approach one uses, the model of prediction is expected to be robust, accurate, and reliable (Sadia et al., 2019). It is also expected to take into account all the variables that might affect the stock's value and performance.
Kimoto et al.  proposed a prediction system of ANNs for predicting stocks listed on the Tokyo Stock Exchange and achieved excellent profits in a simulation exercise. A buying and selling alert system proposed by Tsang et al.  using ANNs to predict stock prices in the Hong Kong Stock Exchange hit an overall ratio of over 70%. Also, studies conducted by [ KIM, T. and ⨯ HA, Y.K., 2019] find that Artificial neural networks (ANNs) can detect nonlinear relationships in the characteristics of data.
Das, Mokashiand Culkin (2018) used neural networks to forecast the movement of the S&P 500 Index and found that the future direction of the index can be predicted by prior movements of the underlying stocks in the index.
Sadia et al., (2019) used the Random Forest algorithm and found it to be a flexible and easy to use machine learning algorithm and that achieves high accuracy rates in forecasting.
French et al.  recommended using the generalized autoregressive conditional heteroscedasticity (GARCH) model to forecast stock prices using the relationship between a stock’s return and its volatility [ KIM, T. and ⨯ HA, Y.K., 2019]
Fama and French (1992) proposed two-factor models using size and book-to-market equity, both of which utilize a company’s fundamental information, to forecast stock prices.
Cao and Tay [ 6, 7] applied support vector machine (SVM) in financial forecasting and compared it with the multilayer back-propagation (BP) neural network and the regularized radial basis function (RBF) neural network;(as cited in SHAO, X., WU, K. and LIAO, B., 2012)
Pai and Lin [ 9] invested a hybrid ARIMA and support vector machines model in stock price forecasting;(as cited in SHAO, X., WU, K. and LIAO, B., 2012) while Kwon and Moon [ 11] proposed a hybrid neurogenetic system for stock trading (as cited in SHAO, X., WU, K. and LIAO, B., 2012);
Jiang and He [ 13] introduced local grey SVR (LG-SVR) integrated grey relational grade with local SVR for financial time series forecasting (as cited in SHAO, X., WU, K. and LIAO, B., 2012)
Hegazy, Osmanand Abdul Salam (2013) proposed a machine learning algorithm which integrates Particle swarm optimization (PSO), and Least-square support vector machine (LS-SVM).
Due to the vast number of options available (Sadia et al., 2019), determining the methods to use in our model was not easy. We considered among other factors the ease of use of the method, popularity, reliability, and expected forecasting accuracy to arrive at the choices we made.
Our choice of the Moving Average was inspired by Majaski (2019) who found this method to be one of the most popular forms of technical analysis and Moving Average (2009) who asserted that it is a very useful for smoothing out volatility and for observing longer-term trends.
Also, our choice of the autoregressive integrated moving average (ARIMA) was guided by Tsai et al. (2018) who noted that the most well-known conventional time series forecasting approach is the ARIMA, and findings by Ariyo et al. (2014) that reveal that ARIMA models do compete reasonably well against emerging forecasting techniques that are utilized today for short term prediction (as cited in Shah, Isah, & Zulkernine 2019).
Our academic experts are ready and waiting to assist with any writing project you may have. From simple essay plans, through to full dissertations, you can guarantee we have a service perfectly matched to your needs.View our services
Equally, experimental results conducted by Nelson et al.  found the LSTM to be more accurate than other machine learning models, such as multilayer perceptron, random forest, and pseudo-random models(SHAO, X., WU, K. and LIAO, B., 2012). This was complemented by studies conducted by [KIM, T. and ⨯ HA, Y.K., 2019] whose findings indicate that the long short-term memory (LSTM) was superior for learning temporal patterns.
Finally, our choice of the RMSE was influenced by Katariya & Jain (2019) who found that the use of RMSE is very common and that it makes an excellent general purpose error metric for numerical predictions. In addition to RSME, we will use the Mean Percentage Error (MPE) and Correlation between Actual and Forecast to measure the performance of our model.
- Abdolmohammadi, M., & Sultan, J. (2002). Ethical reasoning and the use of insider information in stock trading: JBE JBE. Journal of Business Ethics, 37(2), 165-173. doi:http://dx.doi.org/10.1023/A:1015083023298
- Katariya, S. and Jain, S. (2019). Stock Price Trend Forecasting using Supervised Learning. [online] Slideshare.net. Available at: https://www.slideshare.net/SharvilKatariya/stock-price-trend-forecasting-using-supervised-learning [Accessed 7 Nov. 2019].
- KIM, T. and ⨯ HA, Y.K., 2019. Forecasting stock prices with a feature fusion LSTM-CNN model using different representations of the same data. PLoS One, 14(2),.
- Majaski, C. (2019). The Difference Between Fundamental vs. Technical Analysis?. [online] Investopedia. Available at: https://www.investopedia.com/ask/answers/difference-between-fundamental-and-technical-analysis/ [Accessed 8 Nov. 2019].
- Moving average. (2009). FT.Com, Retrieved from https://search.proquest.com/docview/229253194?accountid=196966
- Myer, R. (2011, Nov 22). Moving averages are a main signal on the track: TECHNICAL ANALYSIS. The Age Retrieved from https://search.proquest.com/docview/905154622?accountid=196966
- Prasanna, S. and Ezhilmaran, D. (2016). A Survey of Stock Price Prediction & Estimation Using Data Mining Techniques. [online] Google.com. Available at: https://scholar.google.com/scholar?cluster=15932040444007463065&hl=en&as_sdt=0,14&sciodt=0,14 [Accessed 7 Nov. 2019].
- Sadia, K., Sharma, A., Paul, A., Padhi, S. and Sanyal, S. (2019). Volume-8 Issue-4 | Stock Market Prediction Using Machine Learning Algorithms. [online] International Journal of Engineering and Advanced Technology(TM). Available at: https://www.ijeat.org/download/volume-8-issue-4/ [Accessed 7 Nov. 2019].
- Shah, D., Isah, H., & Zulkernine, F. (2019). Stock Market Analysis: A Review and Taxonomy of Prediction Techniques. International Journal of Financial Studies, 7(2), 26. MDPI AG. Retrieved from http://dx.doi.org/10.3390/ijfs7020026
- Shen, Shunrong, Haomiao Jiang, and Tongda Zhang. 2012. Stock Market Forecasting Using Machine Learning Algorithms. Stanford: Department of Electrical Engineering, Stanford University, pp. 1–5.
- Tsai, M., Cheng, C., Tsai, M. and Shiu, H. (2018). Forecasting leading industry stock prices based on a hybrid time-series forecast model. [online] PLoS One. 2018; 13(12): e0209922. Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6312251/ [Accessed 9 Nov. 2019].
- Nelson DMQ, Pereira ACM, De Oliveira RA. Stock market’s price movement prediction with LSTM neural networks. In: Proceedings of the International Joint Conference on Neural Networks. 2017. p. 1419–26.
- Kimoto T, Asakawa K, Yoda M, Takeoka M. Stock market prediction system with modular neural networks. In: 1990 IJCNN International Joint Conference on Neural Networks [Internet]. 1990. p. 1–6 vol.1. http://ieeexplore.ieee.org/document/5726498/
- Fama E, French K. The cross-section of expected stock returns. J Finance [Internet]. 1992;47(2):427–65. Available from: http://www.jstor.org/stable/2329112
- Tsang PM, Kwok P, Choy SO, Kwan R, Ng SC, Mak J, et al. Design and implementation of NN5 for Hong Kong stock price forecasting. Eng Appl Artif Intell [Internet]. 2007;20(4):453–61. Available from: http://linkinghub.elsevier.com/retrieve/pii/S095219760600162X
- Ariyo A a., Adewumi AO, Ayo CK. Stock Price Prediction Using the ARIMA Model. In: 2014 UKSim-AMSS 16th International Conference on Computer Modelling and Simulation Stock [Internet]. 2014. p. 106–12. http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=7046047
- SHAO, X., WU, K. and LIAO, B., 2012. cursive l]p -Norm Multikernel Learning Approach for Stock Market Price Forecasting. Computational Intelligence and Neuroscience : CIN, 2012, pp. 601296.
- Hegazy, Osman & Soliman, Omar S. & Abdul Salam, Mustafa. (2013). A Machine Learning Model for Stock Market Prediction. International Journal of Computer Science and Telecommunications. 4. 17-23.
Cite This Work
To export a reference to this article please select a referencing stye below:
Related ServicesView all
DMCA / Removal Request
If you are the original writer of this essay and no longer wish to have your work published on UKEssays.com then please: