Multi-Classifier Analysis of Leukemia Gene Expression From Curated Microarray Database (CuMiDa)

In cancer-related investigations, machine learning (ML) techniques are increasingly being utilized to extract gene expression data from microarray experiments. Machine learning is generally recognized to be gaining popularity in cancer biomedical research. For benchmarking and testing such techniques, there are no curated archives that provide high-quality data sets for benchmarking. There are 64 microarray datasets included in the CuMiDa collection, which represents five classes of leukemia gene expression out of a total of 22284 genes. To exclude probes that were erroneous, normalized data and sample quality were evaluated. Support Vector Machine, K-Nearest Neighbor and Random Forest were used to assess primary microarray methods, as were Decision Tree (DT) and Logistic Regression (LR). CuMiDa, a benchmarking and testing tool for machine learning algorithms, was created.