An artificial neural network for pattern classification and visualization
The industrial revolution and the birth of computers has led to a deeper exploration of Artificial Neural Network (ANN), where scientist tries to emulate the biological neural network. Today, ANN has proven to be able to im itate the human neural network and perform task s...
| Main Author: | |
|---|---|
| Format: | Final Year Project Report / IMRAD |
| Language: | English |
| Published: |
Universiti Malaysia Sarawak, UNIMAS
2010
|
| Subjects: | |
| Online Access: | http://ir.unimas.my/id/eprint/6773/ http://ir.unimas.my/id/eprint/6773/4/Sandra.pdf |
| Summary: | The industrial revolution and the birth of computers has led to a deeper
exploration of Artificial Neural Network (ANN), where scientist tries to emulate
the biological neural network. Today, ANN has proven to be able to im
itate the
human neural network and perform task such as solving real world problems.
This study aims to explore in detail an ANN model that is able to perform the
task of pattern classification
and visualisation
, and as well as to evaluate the
performance
of this model. This proposed model is know as the Self
-
Organizing
Map (SOM) or Kohonen’s Map. In order
to
determine it’s accuracy, the SOM
classifier
is tested using a few simulated Gaussian data sets and a real world data
set, the Pima Indians Diabetes da
ta set. The experiments conducted showed that
SOM classifier is able to perform the task of classification and
visualisation
.
However, the classifier obtained a non
-
impressive accuracy of 73.16% in
classifying the real world problem. This results indicate
that SOM is not reliable
compared to other classification applications in literature and can be enhanced in
terms of it’s performance. |
|---|