Predicting Breast Cancer Survivability Using Data Mining Techniques

Last edited by

Quratt ul ain Siddique

Summary:

This paper presents the prediction of the survivability rate of cancer using data mining techniques. In this paper scientists investigated three data mining like Naïve Bayes, the back-propagated neural network, and the C4.5 decision tree algorithms. C4.5 decision tree algorithm is considered to best from remaining two methods. In this study SEER data is used and introduced a pre-classification approach that take into account three variables: Survival Time Recode (STR), Vital Status Recode (VSR), and Cause of Death (COD).

In this paper three data mining techniques are used to find which one is best to find breast cancer survivability rate. In this research Weka toolkit is used for experimentation that used three data mining algorithms in this research raw SEER data is get before processing using different tools.

In this approach missing information from SEER data is completely ignored and included three approaches like STR, VSR, COD. In this study three data mining techniques is compared. The goal is to attain high precision and accuracy from these techniques. These are actually the matrices which are mostly used for the retrieval of the information but there they are considered to be related to the other existing metrics such as specificity and sensitivity.

This paper discussed and resolved the issues, algorithms and techniques and problems related to predict Breast Cancer Survivability using SEER database. It also discussed that among three data minig techniques the C4.5 decision tree is considered to be the best because it shown maximum accuracy ,precision and recall metrics.