Data Mining Techniques for Efficient Detection of Cancerous Masses in Mammogram

Main Article Content

S.Pitchumani Angayarkanni
Dr.Nadira Banu Kamal

Abstract

Breast cancer is the most common form of cancer in women. Early diagnosis of cancerous masses and its size for treatment can prolong the life time of the patients. An intelligent computer-aided diagnosis system can be very helpful for radiologist in detecting and diagnosing micro calcifications patterns earlier and faster than typical screening programs. A number of quantitative models including linear discriminate analysis, logistic regression, k nearest neighbor, kernel density, recursive partitioning, and neural networks are being used in medical diagnostic support systems to assist human decision-makers in disease diagnosis. This research mainly focuses on the decision accuracy of online Back Propagation Neural network model with data mining technique for the diagnosis of cancerous masses from the non cancerous region with accuracy and speed. Conditions where a hierarchical neural network model can increase diagnostic accuracy by partitioning the decision domain into subtasks that are easier to learn are specifically addressed in this paper through decision tree induction method. Self-organizing maps (SOM) are used to portray the 9 feature variables in a two dimensional plot that maintains topological ordering. The SOM identifies five inconsistent cases that are likely sources of error for the quantitative decision models; the lower bound for the diagnostic decision error based on five errors is 0.266. The traditional application of the quantitative models cited above results in diagnostic error levels substantially greater than this target level. A multilayered feed forward neural network is designed for detect. The second stage mixture-of-experts neural network learns a subtask of the automatic detection of diagnostic decision, the discrimination between benign, malignant and normal cases. The diagnostic accuracy of the multilayered feedforward neural network approaches the target performance established from the SOM with an error rate of 0.0012 and accuracy of 99.9%.

Keywords: Preprocessing, Gabor Filter, Decision Tree Induction, SOM and ANN

Downloads

Download data is not yet available.

Article Details

Section
Articles