摘要
Leukemia is blood cancer, including bone marrow and lymphatictissues, typically involving white blood cells. Leukemia produces an abnormalamount of white blood cells compared to normal blood. Deoxyribonucleicacid (DNA) microarrays provide reliable medical diagnostic services to helpmore patients find the proposed treatment for infections. DNA microarraysare also known as biochips that consist of microscopic DNA spots attachedto a solid glass surface. Currently, it is difficult to classify cancers usingmicroarray data. Nearly many data mining techniques have failed becauseof the small sample size, which has become more critical for organizations.However, they are not highly effective in improving results and are frequently employed by doctors for cancer diagnosis. This study proposes a novelmethod using machine learning algorithms based on microarrays of leukemiaGSE9476 cells. The main aim was to predict the initial leukemia disease.Machine learning algorithms such as decision tree (DT), naive bayes (NB),random forest (RF), gradient boosting machine (GBM), linear regression(LinR), support vector machine (SVM), and novel approach based on thecombination of Logistic Regression (LR), DT and SVM named as ensembleLDSVM model. The k-fold cross-validation and grid search optimizationmethods were used with the LDSVM model to classify leukemia in patientsand comparatively analyze their impacts. The proposed approach evaluatedbetter accuracy, precision, recall, and f1 scores than the other algorithms.Furthermore, the results were relatively assessed, which showed LDSVMperformance. This study aims to successfully predict leukemia in patientsand enhance prediction accuracy in minimum time. Moreover, a Syntheticminority oversampling technique (SMOTE) and Principal compenent analysis(PCA) approaches were implemented. This makes the records generalized andevaluates the outcomes well. PCA reduces the feature count without losing anyinformation and deals with class imbalanced datasets, as well as faster modelexecution along with less computation cost. In this study, a novel processwas used to reduce the column results to develop a faster and more rapidexperiment execution.
