Two-Step Verifications for Multi-instance Features Selection: A Machine Learning Approach

摘要: Multi-instance features measurement is an important step in identifying characteristics that are bound to various experimental events. In biological data processing, a set of critical factors responsible for several diseases. Computational simulation will help design optimal tool cost-effective drug design. this regard, the processing big valuable efficient simulation. Recent results generate huge amounts related data. current work, noisy have been treated with three filtering techniques: cross-validated committees (CVCF), iterative partitioning (IPF) and ensemble (EF). A comparison was made these approaches. The filtered datasets were normalized. repeated application normalization techniques removed irregularities structured datasets. Wide ranges comparisons among techniques. After being appropriately structured, normalized transformed accordingly different transformation processes: rank transformation, nominal binary Box-Cox transformation. To prevent false positive negative outcomes experiments, certain key aspects considered: accuracy, sensitivity F-measures. Accuracy experiments relates level precise detection factors; specificity allows selection dominant F-measures ratio between training testing Detailed analysis included study four classifiers deoxyribonucleic acid (DNA) dataset.