CONSTRAINED DIMENSIONALITY REDUCTION USING A MIXED NORM PENALTY WITH NEURAL NETWORKS

CONSTRAINED DIMENSIONALITY REDUCTION USING A MIXED NORM PENALTY WITH NEURAL NETWORKS

                

 ABSTRACT

                              An artificial neural network (ANN), usually called "neural network" (NN), is a mathematical model or computational model that tries to simulate the structure and/or functional aspects of biological neural networks.

                   It consists of an interconnected group of artificial neurons and processes information using a connectionist approach to computation. In most cases an ANN is an adaptive system that changes its structure based on external or internal information that flows through the network during the learning phase.                      Modern neural networks are non-linear statistical data modeling tools. They are usually used to model complex relationships between inputs and outputs or to find patterns in data.

                    For our work, neural networks provide an ideal platform, since they can be trained to provide near optimum classification for almost any problem.

                  

                   We use the neural net here because of the straightforward relationship between the functional dimensionality of the problem and the number of neurons in the first hidden layer of the net. The result that is obtained May also be used directly on the pruning problem of neural networks to reduce the total number of neurons.

Existing System:                    

                   Existing system does not have any formulae to neutralize the unauthorized access of the tested platform. There is no record of to identify whether the pan card holder is taken the share or it is used by others as BROKER.

It is having more chance for the BROKERS do destabilize the system. The existing neural does not implements fixed dimensionality for neural layers.

Disadvantage of Existing System

                   Existing system uses fixed norm algorithms which operate for the system that has predefined inputs, for change in input needs the algorithm to be restructured which makes it more complex and time consuming.

The existing neural does not implements fixed dimensionality for neural layers.

Advantage of Existing System

                   Neural network requires less formal statistical training for the hidden layer neurons. Ability to detect all possible interactions between predictor variables, and the availability of multiple training algorithms.

                   The input object will be transformed automatically into its binary forms on the computational process.

 Proposed System

           Our system tasted with the day to day STOCK EXCHANGE system, in current system there is no provision of identifying the user that taking share and the source of money that comes from which opens our system more open towards black marketers or black money may be within or outside country.

         

          To protect we have done a quite efficient system that identifies the user based on their user image, here to purchase or sale any share from the STOCK EXCHANGE user needs a strong authentication which cannot be avoided.

         

          To make our system more real we implemented our system with latest technologies which having a better market perspective such as Struts 2.X, Hibernate 3.X of J2EE technology.

         

          Our solution towards the problem that we are facing in existing system cannot be overlooked.

                              

SYSTEM REQUIREMENTS

Hardware Specification:

•         Processor             : Dual Core/Pentium.

•         Hard Disk            : Min 20GB.

•        RAM                    : 512 MB.

Software Specification:

•         Operating system           : Windows XP Professional.

•         Front End             : J2EE

•         Back End             : MySQL

•         Framework                    : Struts Hibernate.