Model Comparison for Breast Cancer Prognosis Based on Clinical Data

Abstract

We compared the performance of several prediction techniques for breast cancer prognosis, based on AU-ROC performance (Area Under ROC) for different prognosis periods. The analyzed dataset contained 1,981 patients and from an initial 25 variables, the 11 most common clinical predictors were retained. We compared eight models from a wide spectrum of predictive models, namely; Generalized Linear Model (GLM), GLM-Net, Partial Least Square (PLS), Support Vector Machines (SVM), Random Forests (RF), Neural Networks, k-Nearest Neighbors (k-NN) and Boosted Trees. In order to compare these models, paired t-test was applied on the model performance differences obtained from data resampling. Random Forests, Boosted Trees, Partial Least Square and GLMNet have superior overall performance, however they are only slightly higher than the other models. The comparative analysis also allowed us to define a relative variable importance as the average of variable importance from the different models. Two sets of variables are identified from this analysis. The first includes number of positive lymph nodes, tumor size, cancer grade and estrogen receptor, all has an important influence on model predictability. The second set incudes variables related to histological parameters and treatment types. The short term vs long term contribution of the clinical variables are also analyzed from the comparative models. From the various cancer treatment plans, the combination of Chemo/Radio therapy leads to the largest impact on cancer prognosis.

Breast Cancer Diagnosis Using a Hybrid Genetic Algorithm for Feature Selection Based on Mutual Information

“…It turns out that there is no accurate definition of "Genetic Algorithms" formally accepted by the evolutionary computation community [14]. However, it can be said that Genetic Algorithms are adaptive heuristic search algorithm which are invented by Holland in the 1960s and inspired from Darwin's 1 Available from: https://archive.ics.uci.edu/ml/datasets.html theory of evolution "survival of the fittest" [15]. Five important factors can change how the optimization scheme is performed by Genetic Algorithms (GA): population creation, fitness function, selection schema, genetic operators and stopping criteria.…”

“…Breast cancer is the most common cancer in women in both developed and developing countries where the number of breast cancer cases worldwide was estimated at 14.1 million new cases and 8.2 million deaths in 2012 [1]. Successful early detection can bring better treatments to patients, and can help medical experts make important decisions about patient healthcare.…”

“…The selected subset of features are then input in two classifiers: the Support Vector Machine (SVM), and the k-Nearest Neighbor (k-NN) classifier for performing the prediction task. The experiments were performed on the Wisconsin Diagnostic and Prognostic Breast Cancer dataset 1 . The results show that only a subset of the features were required for reaching highest classification accuracy.…”

Developing a periodontal disease antibody array for the prediction of severe periodontal disease using machine learning classifiers

“…Since machine learning techniques are increasingly being applied in medical diagnosis, we therefore applied five classification models and compared their diagnostic performance in distinguishing NOR from SP patients. These five models included support vector machine (SVM), random forest (RF), k nearest neighbor (kNN), linear discriminant analysis (LDA), and Classification and Regression Trees (CART) . ‒ This study represents the first comparison of classification models of periodontitis using GCF.…”

“…These five models included support vector machine (SVM), random forest (RF), k nearest neighbor (kNN), linear discriminant analysis (LDA), and Classification and Regression Trees (CART). [28][29][30][31] This study represents the first comparison of classification models of periodontitis using GCF.…”