Dear Dr. Amzallag,

                                 I have published some research relating to coal mining and coal stock markets in which I have considered the three dimensional efficiency problem viz. physical efficiency*economic efficiency*economic efficiency with optimal growth genetic algorithms, which are supposed to violate Haag's Theorem and Quantum Mechanics for example but I have established market potential equilibrium using D-Branes -String Duality (please refer to my Information Structure on my RG page) by using the Econophysics Expectations operator which I have proved in my earlier research (please refer to my RG page) satisfies Von-Neumann actions and allows expressing Quantum Mechanical systems in terms of probabilities using the Weak Law of Large Numbers (Space Time interaction picture according to Schroedinger) to conform with coal sector stock markets experiments (systemic experiments). Here as I have already stated human pshycology factor through the partly satiable (bounded rational) money-energy conversion system drives the Lagrangean actions which is the basis for the Von-Neumann actions. SKM QC

Thare are efforts in this direction eg R. L. Peterson and his company marketpsych

Hello Sir,

my area of work is that, my latest work is on comparative predictive modelling on CNX Nifty with the help of Artificial Neural Network. I could establish in that paper that prediction could be possible with more than 98% accuracy if the neuron layers are chosen correctly. Since Indian capital markets are week form of efficiency and human decisions can be predicted well with ANN, so I have used it to good effect. However that paper is going out in October 2015. Also I am working on composite Index blending human emotions and stock market quantification. That is the core of my Thesis. Unfortunately that is also under process. May be in another six months I will be able to communicate with you with more value proposition.

Regards

Dear Soumitra,

Thank you for your answer.

Your line of research is very innovative and interesting. Your answer being by necessity condensed, I have not grasped all the techniques derived from physics that you use but I will look at your papers and  see if it could help me in my angle of research.

Dear Christos,

Thank you for your referral. Marketpsych is a very good example of the growing importance of behavioural finance principles in stock market investments decisions.

I believe that they promote the use of these principles by human (as opposed to algorithm driven computers) investors. Nevertheless, it is quite relevant to my research. Thank you

Dear Bikramaditya,

Your line of research is very interesting and could lead to some interesting exchanges with mine.

To be more precise, I am concentrating on the US stock market because it is the most influenced by high frequency trading, currently around 50% of the volume. There is a clear trend for human investors to use behavioural finance principles as a guide to investments. At the same times the influence of algorithm driven computer trading is dominant, hence my question as it seems that the two trends can only be compatible if some psychological factors are introduced in the algorithms. If they are, what are they?

Thank you 

Hello Sir,

thanks for your reply.

In this case specific to your query

1. Ethnic biases calculation Mechanism(ancient wisdom of investment)

2. Greed Calculation Mechanism

3. Fear calculation Mechanism

4. Event effect calculation Mechanism(Policy Rate Announcements etc.)

These psychological characters could be able to make it more accurate in nature according to my limited knowledge base.

Warm Regards

Bikram

 A study of the current authors in AT/HFT literature shows there is considerable bent in the quant field to detect MLP based /Wavelet based, non technical factors for stock selection. Also the complex implementations of NN effectively match 'liquidity trading' by machines to discover/snipe human price setters. GA mechanisms rely on the original HP/IBM argument with trading algorithms and NN rely on training using index/market wide data. 

My literature review and/or a perusal of the latest (2012-) journals in Quantitative Finance and other science direct journals is likely to be the most fruitful. 

Dear Amit,

Thank you for your very interesting comment.

. I have red some literature on wavelet based systems enhanced by bayesian time series modeling using NN. 

These systems improve a lot on classic modeling and, as you mentionned, are meant to detect human price setting mostly in the short term.However, they are based on past price series.

My question refers to the principles of behavioural finance such as anchoring, loss aversion or gamblers bias. I wonder if these principles that govern a substantial amount of human trading could be (or already have been) introduced in trading algorithms giving machines a new predictive tool rather than relying exclusively on past series.

Herding behavior influences stock market direction too.

I agree with Amit Mittal