- Laptop or Computer
- University of Wisconsin Breast Cancer Data Set
- Eclipse Indigo
- Java Runtime Environment
- Google App Engine
- Google Web Toolkit
- Each neural network should adhere to the following
- Each trial
shall use a different set of randomly selected instances for testing.
commercial package shall be optimized to yield best results. For example,
if a package can support multiple hidden layers, the option will be evaluated to determine the best settings. The
custom network will be implemented
with proper settings.
neural network will use the University of Wisconsin Original Breast Cancer Database.
- Trials of
the same implementation will all be run using the same settings.
trials will be run on the same computer.
- The number
of iterations for custom neural networks will be 1,500
of diagnostic success achieved by the networks.
Abridged versions of the procedures for this project are
Experiment Phase 1
- Understand significance of each of 9 inputs:
clump thickness, uniformity of cell size, uniformity of cell shape, marginal
adhesion, single epithelial cell size, bare nuclei, bland chromatin, normal
nucleoli, and marginal adhesion.
- Architect and optimize a neural network model for
each of three commercial softwares, network should adhere to the following control
- Reserve 70 instances for testing. In other
words, those instances should not
be used for training.
- Each trial shall use a different set of randomly
selected instances for testing.
- Each package shall be optimized to yield best
results. For example, if a package can
support multiple hidden layers, those results should be evaluated to decide
- If multiple neural network training
architectures are supported, a back propagation technique shall be chosen.
- Execute 10 trials for each network
implementation and capture number of properly predicted malignant and benign
tumors and also number of false malignant and false benign predictions.
- Analyze results to determine success and
failures of implementations.
- Determine if modern neural networks were
successful at predicting malignant versus benign when including outliers.
- Identify areas for improvement.
Experiment Phase 2
pseudocode for a custom developed breast cancer neural network, including the
following algorithmic components.
Input Layer. Convert inputs to binary
inputs to simulate the on/off firing of neurons.
Function. A logistic function that removes the linearity from processing.
Function. Matrix math function to propagate
neural firings through the network.
Function. Incorporate malignancy
Assessment. Evaluate using multiple independently
- Define a neural network model with artificial input layer.
a way to weight malignant false negatives higher.
a custom neural network in Java.
logic that allows the network to rule masses inconclusive.
multiple base networks using the data from the dataset, allowing the computer
to do all weighting on its own.
- Tune the
network with different number of hidden nodes and malignancy weightings to
identify optimal configuration.
- Test the
network by training the network with all samples but one. Run 10 different
trials for each sample (This should result in 6,800 trials).
network using all samples in the training set to compare results and determine
a website to host the neural network implementation and provide an interface for
- Design and implement a web service suitable
for integration with the cloud via Google’s AppEngine.
- Deploy the web service and web application to
- Test the network with different test sample
sizes to determine correlation to success rates. Use a large number of tests to reduce the
impact of randomness. Running 1,500
networks for each training size at 20 increments will result in over 7,000,000
- Analyze results and present findings.