This work addresses the issue of variable selection within the context of breast cancer classification with mammography. A comprehensive repository of feature vectors was used including a hybrid subset gathering image-based and clinical features. It aimed to gather experimental evidence of variable selection in terms of cardinality, type and find a classification scheme that provides the best performance over the Area Under Receiver Operating Characteristics Curve (<i>AUC</i>) scores using the ranked features subset. We evaluated and classified a total of 300 subsets of features formed by the application of <i>Chi-Square Discretization, Information-Gain, One-Rule</i> and <i>RELIEF</i> methods in association with Feed-Forward Backpropagation Neural Network (<i>FFBP</i>), Support Vector Machine (<i>SVM</i>) and Decision Tree J48 (<i>DTJ48</i>) Machine Learning Algorithms (<i>MLA</i>) for a comparative performance evaluation based on <i>AUC</i> scores. A variable selection analysis was performed for <i>Single-View Ranking</i> and <i>Multi-View Ranking</i> groups of features. Features subsets representing Microcalcifications (<i>MCs</i>), Masses and both <i>MCs</i> and Masses lesions achieved <i>AUC</i> scores of 0.91, 0.954 and 0.934 respectively. Experimental evidence demonstrated that classification performance was improved by combining image-based and clinical features. The most important clinical and image-based features were <i>StromaDistortion</i> and <i>Circularity</i> respectively. Other less important but worth to use due to its consistency were <i>Contrast, Perimeter, Microcalcification, Correlation</i> and <i>Elongation</i>.