Problems in the Stock Market

Problems in the Stock Market

Background

Business operators in US have one thing to be sure of; that investing in US stock market has never been save under any circumstances. This may be attributed to the fact that investment in equities does not guarantee 100% outcome.  A lot more people have come to think of investing as a game of probabilities. This may be attributed to the many happenings that occur in the stock market and which significantly affect the investment in either a negative or positive way.

The stock market has become a trial and error for investors especially those in the middle class. The enormous number of businessmen from all categories has made the stock market becoming more volatile and less predictable investment sector than ever. The trend has made even the best designed investments to fail. The unpredictability trend has also made even the best brokers to fail in providing appropriate advice.  The external shocks have been among the factor causing unpredictability in the stock market.  From time immemorial, some of these shocks have included inflation, political instability, oil and fuel price increase, defaults by major firms and the high rate of interest. These and other shocks are not easily predicated, and apparently, their effects on the stock market are not easily predicated either (Barrie, 188).

Scatigna, (2013) observes that forecasting the financial markets in the coming years is challenging and uphill task.  This due to the many variables that challenges the outcome,  Other factors influencing this  phenomena    include the geographical and  political  events  that largely dictates the  course  of financial markets in the coming years.  The news media has also posed as a problem in forecasting financial markets in the recent times.  These markets are captive to news which sways them down or up.  Unless one has a capability of seeing future events, predicting the performance of any financial markets confidently is a difficult venture.

 

The basic question that has been pondered by many investors is on how to trade in the stock market in an efficient way.   This has been an equally challenging phenomenon as ever before. Apparently, conventional strategies such as Markowitz’ portfolio of mean variance when applied with parameters that have been estimated from data are poised to give a remarkable impulsive portfolio weights. This is basically due to the complexity of estimating the anticipated returns in an accurate manner (Bielecki,et al, 2005, 214).

 

Trading strategies for earning abnormal returns may be developed by following signals of organizational distress or recovery. By use of signals developed by two famous bankruptcy models, that is the Wilcox X value model  and the  Altman Z score, the researcher grouped the NYSE organizations  in accordance to whether  they  were moving from positive  gains to negative gains. For a period of 15 months, before the issuance of the yearly report that influenced and alteration in the state, organizations classified by the Altman model as recovering from distress indicated significant abnormal positive returns. On the other hand, those that were classified as deteriorating portrayed negative returns.  Both groups in this perspective went on exhibiting abnormal returns in the anticipated returns over nine months prior to the announcement date. By use of the Wilcox model as the discriminator of recovery, the researchers established no abnormal return behavior prior or after the distress forecasting but substantial expectation of recovery signs.

 

It is generally acknowledged that prediction of the stock market is a cumbersome task.  It is therefore with this reason that intelligent systems for predicting the stock market have widely been developed. Some of these systems include genetic programming and nero-genetic among others. The systems behave more like a process of random walk and vary in time. The complexity of the stock market opens away for the significance of the intelligence prediction systems. In then last ten years, stocks and future investors have come to rely extensively on these intelligence decisions in making their business deliberations. One of the goals in this study is to examine the significance of genetic programming and neural genetic networks in the prediction of the stock markets.

 

Genetic programming

Genetic programming is a newer development in the field of evolutionary algorithms. This system extends to the conventional genetic algorithms by accepting the dispensation of non linear structures (Koza, 167).  Holland (1975) defines a genetic algorithm as a randomized search procedure that works on a group of individuals or solutions that are encoded as linear bit of strings. This population develops over time through the application of operators which resemble those in nature that is the selection of fittest, crossover and mutations. GP is an evolutionary computational method that is most based on what is referred to as “tree representation”. This type of representation is tremendously flexible since trees can better present much of mathematical equations, computer programs, or full models of the systems processes (127). Koza continues to explain that genetic programming is used to produce a model that is based on a one day ahead in predicting stock exchanges. In yielding a higher rate of return on the investment, the model is tried for approximately fifty business days consecutively in yielding high returns on investment.

Kelly et all 2004 observes that in the recent perspectives, genetic programming has proved to be  a highly effective and significant method in identifying  data relationship for which there is a lack of a more precise  and theoretical construct. The GP search could be equally useful in developing business strategies that are based on the agent economic models. These strategies employ prices in the stock market and technical indicators, for instance the moving average divergence or convergence and a diversity of exponentially weighted poignant averages in generating, selling and buying signals.

Genetic programming solves stock market problems by imitating the evolution process (Koza, 3). Under this system, a population of   candidate solutions is retained and customized   in a manner that weak solutions are removed and replaced   by strong variants until minimal or no progress is achieved with regard to the best solution in population. When this happens, it is said that the problem has converged to a solution that is feasible (Karjalainen, 245).

Another significant use for this system is in predicting financial asset prices. It completes the first phase in the construction of technical trading. Technical strategies can now be generated on the framework of simulated stock market data. It is expected that the technical trading agents such as investors play an important role in assisting the model to imitate the real global financial markets. Further, this model can be very useful in developing strategies that are aimed at mitigating the negative effects. Effectual execution may take a few iterations of developing and testing technical investors. This is because the inclusion of the simulation is poised to alter the behavior of the replicated stock market prices (Becker, 26)

The fitness function is another significant element in the genetic program.  This component measures the achievement of a solution strategy.  Since the objective of the agents is to make better use on the investment returns, the fitness function is a common element in computing these returns in the sense that one dollar would achieve by using the strategy of the tree.  For the fitness function, every tree is assessed at every time step.  If the tree happens to return a value which is greater than 0, a buy signal is developed. If the value happens to be lower than 0, then a sell signal is developed.  If this tree happened to be in the market at a previous moment and a buy signal was realized, then there is no action to be considered. Consequently, a sell signal has no influence if the tree does not appear in the market. At the beginning of the training, every tree starts with a dollar. When the tree sells, the first stock is multiplied by the ration of the buying price to selling price.  After it runs through the times, the original dollar is subtracted off for the fitness to counter the trading returns. Further, it should be noted that “in the market” translates to the investor’s money while “out of the market” translates to the withdrawal of the investor’s money (Neely, 111).

Nero-Genetic (ANN)

Studies have pointed out that neural networks have facilitated the achievement of better results in the prediction of the stock market in comparison to either linear or non linear functional form to model the movement of prices in the stock market.  Varahrami, (2012) observes that Neural networks have the benefit of simulating the models that are non linear in a case of the existence of the slight priori knowledge structure problem or the number of input variables that are immeasurable are much higher or when the systems are characterized by chaos.

Neural genetic networks or simply neural networks are a category of generalized, non linear models triggered by studies involving the human brain.  The main benefit accrued from these systems is that, they have a capability of approximating various aspects of non linear function to a subjective degree of precision with a relevant number of hidden units. These networks obtain their intelligence from the learning procedure. Consequently, this intelligence generates the ability to automatically adapt or associate the memory to perform specific functions (Goldberg, 69).

Artificial Neural Networks (ANN) is a network that is based on biologically stimulation. The method is based on the composition of neurons and the process of making decisions in the human mind (Madala, and Ivakhnenko, 384). Explained differently, Artificial Neural Networks is a mathematical analogue technique of a human nervous system. It is very useful in pattern identification, prediction and pattern sorting processes. This method has been reviewed and proved to be reliable by various scholar and authors especially in the case when the related systems are so complex that the underlying processes or relationships are not fully understandable or harbor chaotic characteristics (Krishnan, 328).  Development of Artificial Neural Networks for any particular systems incorporates three crucial elements a) proper training algorithm b) network topology c) transfer role.  The network learns though adjusting the interconnections between layers.  As the training or learning procedure is finalized, a relevant output is produced at the output layer. The learning process may either be unsupervised or supervised. If the problem is with regard to prediction, then the learning process should be supervised so that the anticipated output is prior assigned to the network (Darvizeh, et al, 87).

 

 

Work Cited

 Barrie, Scoot. “The Complete Idiot’s Guide to Market Timing” Alpha Books, 2003

Scatigna, Lou. “Economic and Financial Market Forecast” 2013. The Financial  Physician.

Bielecki, T.R., Jin, H., Pliska, S.R. and Zhou, X.Y. (Continuous-time meanvarianceportfolio

selection with bankruptcy prohibition. 2005. Math. Finance 15 213–244.

MR2132190

Koza JR. “Genetic programming: on the programming of computers by means of natural

Selection”. Cambridge, MA: MIT Press, 1992.

Holland, JH. “Adaptation in natural andarti3cial systems” Ann Arbor, MI: The University of

Michigan Press, 1975.

Kelly. A.,  Farnsworth,G.  “Successful Technical Trading Agents Using Genetic Programming”

Sandia National Laboratories Albuquerque. 2004.

 

Karjalainen, Risto and  Franklin Allen. “Using genetic algorithms to find technical

 

Trading rules” Journal of Financial Economics, 51:245–271, 1999.

 

 

Becker, Lee  and Mukund Seshadri. “GP-evolved technical trading rules can

 

outperform buy and hold” In Proceedings of the Sixth International Conference

on Computational Intelligence and Natural Computing, Embassy Suites Hotel and

 

Neely Christopher J. “Risk-adjusted, ex ante, optimal technical trading rules in

 

Equity markets” Working Papers 99-015D, Federal Reserve Bank of St. Louis,

1. available at http://ideas.repec.org/p/fip/fedlwp/99-015.html

Varahrami, Vida.   “Good Prediction of Gas Price between MLFF and GMDH Neural Network”

Faculty of Economics, University of Tehran, Tehran, Iran. 2002.

 

Goldberg D. E., Deb K. “A comparative analysis of selection schemes used in genetic

Algorithms, Foundations of Genetic Algorithms” Morgan Kaufmann press, pp. 69-93, 1991

Madala,H.R. and Ivakhnenko,A.G. “Inductive Learning Al-gorithms for Complex Systems”

Modeling; CRC Press Inc., Boca Raton, 1994, p.384.

 

Krishnan, S.M., Srinivasan, N., Ravichandran, V., Chan, K.L., Vidhya, J.R., Ramakirishnan, S.

“Exponentiated backpro-pagation algorithm for multilayer feedforward neural net-work Neural Information Processing” ICONIP apos; 02. Proceedings of the 9th International Conference on Volume 1, Issue , 18-22 Nov. 2002 Page(s): 327 – 331 vol.1

Darvizeh, A., Darvizeh, M., Gharababaei, H. “Modeling of explosive cutting process of plates

using GMDH-type neural network and singular value decomposi-tion” Journal of Materials Processing Technology, 2002, vol. 128, no. 1-3, pp. 80-87, Elsevier Science

 

Last Completed Projects

Are you looking for a similar paper or any other quality academic essay? Then look no further. Our research paper writing service is what you require. Our team of experienced writers is on standby to deliver to you an original paper as per your specified instructions with zero plagiarism guaranteed. This is the perfect way you can prepare your own unique academic paper and score the grades you deserve.

Use the order calculator below and get started! Contact our live support team for any assistance or inquiry.

[order_calculator]