Dear Prof Dr Mahmoud

I am sure that you have working knowledge of ANN toolbox in Matlab. First, you need to acquire correct real time data related to stock market wrt actual stock prices that you want to predict and related parameter which affected the prices. You need to keep stock price as Target and the parameters affecting the trend as Input in the training process.

As far as my knowledge works, there is no specific real time stock market toolbox in Matlab. However, I believe the application can be developed using offline training of ANN and testing on known examples to validate the results first. Later, the adaptability can be incorporated.

Thank you for the prompt response. Two related issues remain. 1) Which technical and fundamental analyses are more appropriate. 2) I want the output ranks among different stocks (not just predict within one share based on historical data).

Short term prediction: technical; long term prediction: fundamental

I think for this u first need to (1) Identify the maximum possible parameters that effect the effect a stock performance

(2) You need to have sufficient data (deterministic) so that your NN model can actually predict stock performance after sufficient training.

(3) I think simple ANN wont work here. You need to incorporate probabilistic model working side by side embedded inside you NN model. The real challenge is how well you model uncertainties.

Please anyone, could you explain what are the parameters we consider for technical and fundamental analysis.

And what are the issues/parameters affect the market

What in my mind is ,

Difference between Trading and Investment :

1. Trading : Daily / Intraday ( what you invested should take back on same day)

Need Technical Analysis

2. Investment : What you invested for long term ( 6 days to any no of days )

Need Fundamental Analysis

In the world of stock markets, I would say short term could be 6 days or less for very liquid stocks (constantly that sells and get buyers in the same day) to a couple of months for less liquid stocks - you need technical analysis. Long term is very hard to define, one student of mine research mining stock and long term there would be 6 months or more.

Technical stocks works analyzing the evolution of price in the short term, max and mins per day, opening prices, closing prices daily.

Fundamental analysis uses forecasting of market growth, debt ratio, dividends payed last couple of years, stuff you get from the financial statements of the company and the market forecasting

Thanks Mr.Glen Rodriguez

I need examples for using neural networks..

My research area is image segmentation and stock data analysis.

So please would you please suggest any video tutorial or

a book to learn neural networks with different examples for various domains.

I have interest to know how to construct a NN to a problem

Keep in mind that when we apply neural networks to science problems we implicitly rely on the laws of nature. As long as the data set used to calibrate and assess the performance of the neural net covers most possible states of the system, the neural net should provide valuable information (if the system changes the neural net will need to be recalibrated). In the stock market, valuations include a subjective weighting of the relative importance of the inputs that will vary over time. Data such as the price and PE history of the stocks along with relevant market indexes and market sectors of the stocks are likely good starts for the predictor set but no matter what group of predictors you select, how the market values their relative importance could change during the time frame of your data set.

To learn how to use neural nets, I also recommend Matlab if you have access to it. The neural network toolbox has now a graphical user interface that makes it pretty easy to implement a simple neural net and start experimenting with model parameters (has worked well for class projects for my 1 semester graduate forecasting class). There are many other ways to get started with neural nets that others may recommend.

I found similar topic with new answers below:

Can Machine Learning Techniques Be Used To Predict Stock Prices?

https://www.researchgate.net/post/Can_Machine_Learning_Techniques_Be_Used_To_Predict_Stock_Prices?cp=re65_x_p3&ch=reg&loginT=KejNagndqT-nDtsTdTpkAn8n-rSxeT6G&pli=1#view=53ad47bfd11b8b224d8b45c1