Neural predictions.

      

 

  		   
  • Neural Networks:
  • Here you are some details of how Neural Network can forecast future moves of stock markets:
  • Neural networks are build to have architectures like the human brain using some advantages of the conventional digital computers. These advantages provide faster and more "intelligent" reasoning. Once trained, they will behave always the same way under identical conditions, against the changing human behavior. Computers are at first programmed to get ability to learn and then are trained by repeated presentation of data in a complex way.
  • Once trained they can interpret new data. There exists a lot of different neural network architectures, we use Radial Basis Function (RBF) nets that offer the fastest training, while retaining optimum performance, and avoiding learn the noise environment typical of financial series of data.
  • These RBF's have a wide range of application, and we have adapted them to forecasting of market indices and predicting commodity trading prices. Other uses are

    •Portfolio management

    •Sales forecasting, stock control and promotions

    •Price elasticity analysis and valuation

    •Operations Management and service provision

    •Marketing analysis

    •Audience prediction and demand forecasting

    •Distribution management, transportation and "people handling"

    •Database marketing: demographic segmentation by lifestyle, income etc

    •Matching product technical attributes with market sectors

    •        
           

     

          		     
           
           
          ... back to Market Forecasting Page