You are on page 1of 4

c 

 
   

 
         
  

 
 
 
    
       
      
    
   
 
 
   
 
 
   
    
  



 
 
  !""# $
% 
&      
#
' 
    (((    
   )  

 
 
 
 
  
       
       $%&
    
 
  
    

  

   
 

Probabilistic neural networks algorithm is one algorithm that can be used to provide a
robust classification of datasets. Probabilistic neural networks works best in learning of
complex real-world sensor data which are generally characterized by non-linear ties, high
dimensionality noisy, imprecise, and imperfect or error prone sensor data, and lack of a
clearly stated mathematical solution or algorithm.

A PNN is an implementation of a statistical algorithm called kernel discriminant analysis in


which the operations are organized into a multilayered feed forward network with 3 layers
of nodes which are input layer, radial basis function layer and competitive layer. The
architecture for this system is shown in figure 3.
O 



In general PNN is one kind of supervised neural network used in classification. It is derived
from Radial Basis Function (RBF) Network. When an input is presented, the Input Layer
computes distances from the input vector to the training input vectors and produces a
vector whose elements indicate how close the input is to a training input. Then the Radial
Basis Layer sums these contributions for each class of inputs to produce a s its net output a
vector of probabilities. Finally, a t Competitive Layer, the compete transfer function on the
output of the second layer picks the maximum of these probabilities, and produces a 1 for
that class and a 0 for the other classes.

PNN is adopted for it has many advantages. Its training speed is many times faster than a
backpropagation network. PNN can approach a Bayes optimal result under certain easily
met conditions.

However, PNN is subjected to some disadvantages. The main drawback of PNN is the need
to transform the high dimensions attributes information using principal component analysis
(PCA) to reduce the dimension ality and it is computationally expensive. In high dimensional
space problem, a non-linear classification algorithm is more suitable to attain a better
classification.
ð
 

 

   

ëp ARTIFICIAL NEURAL NETWORKS AND OTHER METHODS OF IMAGE CLASSIFICATION


ëp Probabilistic Neural NetworkFor Brain Tumor Classification O 
ëp Computational Properties of Probabilistic Neural Networks Grim 2010
ëp Comparative Exudate Classification Using Support Vector Machines and Neural
Networks Alireza 2002
ëp þ   
 
    þ 
  
     
        
   
!"#* $#%
ëp c  
 c          
     

You might also like