Optimized Classification Method for Prognosis of Breast Cancer Using RapidMiner

摘要

The quality of the data is one of the most important factors influencing the performance of any classification or clustering algorithm. The attributes defining the feature space of a given data set can often be inadequate, which make it difficult to discover interesting knowledge or desired output. An electrical impedance measurement in samples of freshly excised tissue from the breast is one of the techniques to discriminate Carcinoma, Fibro-adenoma, Mastopathy, Glandular, Connective and Adipose (cellular transformation). In this context, the state-of-the-art software RapidMiner already provides easy to use interfaces for developing and evaluating classification and validation and Machine Learning applications. C.Y.V.Watanabe reported output at very high accuracy and prove the value of validation using SACMiner. The aim of this work is to implement a RapidMiner extension that contains the most applicable classification algorithms to enable non-experts to easily apply them. Secondly an evaluation of the implemented algorithms was carried out in an attempt to show the relative strength and weakness of the algorithms as Watanabe’s work. Simultaneously attempt was made to use evolutionary parameter optimization for selection of optimal values of model parameters for enhancement of model accuracy. Efforts were also made to identify the attributes that influences the accuracy of the result using weighting attributes. The performance of the implemented and the proposed algorithms were evaluated on real world data sets from the UCI machine learning repository.