High Performance Modeling of Intelligent Pattern Recognition with Enhanced Fault-Tolerance in Real Time
Journal: International Journal of Advanced Computer Research (IJACR) (Vol.4, No. 14)Publication Date: 2014-03-16
Authors : Renukaradhya P.C; Shobha M;
Page : 402-406
Keywords : Artificial Neural Network; Patterns; Forced recognition; Custom recognition; Fault Tolerant.;
Abstract
Designing an ANN which could recognize the learned patterns even if there is variation in applied test patterns from learned patterns. A mechanism has been developed which provided the recognition facility intelligently. Recognition of patterns can be broadly categorized into two classes. When precision of recognition is not defined, term name “Forced recognition” given to the process. When precision of recognition is properly defined termed “Custom recognition” given to process. Analysis of fault tolerant property of feed forward architecture will be given training with back propagation method. Under this, analysis of effect of initially selected random weights and what should be the nature of random weights so that to maximize the fault tolerance capability of system has done. Analysis can be done with two different distribution namely Gaussian distribution and Uniform distribution. Effect of faults at output is also a function of fault position in ANN system like Hidden layer weight, Output layer weights, with processing elements at hidden layer. Analysis capability of back propagation algorithm itself is to tolerate the fault by learning process. A development of test mechanism to check faulty system in coming future is ANN system in hardware world i.e. on the VLSI chip. Once the architecture implemented it is required a mechanism to check the functioning. Analysis of internal parameters of ANN is completely research work with behavior of internal parameters, which will provide all responsible factors behind success of an ANN.
Other Latest Articles
- Smart Test Case Quantifier Using MC/DC Coverage Criterion
- Operating System’s Changes When Installing an Anti-virus Application
- An efficient Optimization method for Data Classification
- A Survey on Different Wake-up Scheduling in WSN
- Hands Free Browser-‘An Interactive Speech Browser for Visually Handicapped’
Last modified: 2014-12-17 14:57:24