Page 196 - DCAP208_Management Support Systems

Unit 11: Neural Networks

11.2.3 Learning Algorithm Selection Notes

After the network structure is chosen, we need to find a learning algorithm to identify a set of
connection weights that best cover the training data and have the best predictive accuracy.
For the feedforward topology we chose for the bankruptcy-prediction problem, a typical approach
is to use the backpropagation algorithm. Because many commercial packages are available on
the market, there is no need to implement the learning algorithm by ourselves. Instead, we can
choose a suitable commercial package to analyze the data.

11.2.4 Network Training

Training of ANN is an iterative process that starts from a random set of weights and gradually
enhances the fitness of the network model and the known data set. The iteration continues until
the error sum is converged to below a preset acceptable level. In the backpropagation algorithm,
two parameters, learning rate and momentum, can be adjusted to control the speed of reaching
a solution. These determine the ratio of the difference between the calculated value and the
actual value of the training cases. Some software packages may have their own parameters in
their learning heuristics to speed up the learning process.

!
Caution It is important to read carefully when using this type of software.

11.2.5 Testing

Recall that in step 2 of the development process shown in Figure 11.9, the available data are
divided into training and testing data sets. When the training has been completed, it is necessary
to test the network. Testing (step 8) examines the performance of the derived network model by
measuring its ability to classify the testing data correctly. Black-box testing (i.e., comparing test
results to historical results) is the primary approach for verifying that inputs produce the
appropriate outputs. Error terms can be used to compare results against known benchmark
methods.

Task Analyze the importance of testing.

11.2.6 Implementation of an ANN

Implementation of an ANN (step 9) often requires interfaces with other computer based
information systems and user training. Ongoing monitoring and feedback to the developers are
recommended for system improvements and long-term success. It is also important to gain the
confidence of users and management early in the deployment to ensure that the system is
accepted and used properly.

Self Assessment

Fill in the blanks:

13. ................... of ANN is an iterative process that starts from a random set of weights and
gradually enhances the fitness of the network model and the known data set.
14. ................... is the primary approach for verifying that inputs produce the appropriate
outputs.

LOVELY PROFESSIONAL UNIVERSITY 189