rMisbeta

rMisbeta imputes missing values and handles outliers using a robust iterative algorithm based on minimum beta divergence, enabling accurate data preprocessing for high-dimensional omics datasets.

Key Features:

• Minimum Beta Divergence Imputation: Applies a robust iterative framework that simultaneously imputes missing values and mitigates outlier effects using minimum beta divergence estimators.
• Robust Performance Evaluation: Benchmarked against Zero, KNN, robust SVD, EM, Random Forest, and Weighted Least Square Approach (WLSA) using metrics including Frobenius norm (FOBN), accuracy (ACC), sensitivity (SN), specificity (SP), positive predictive value (PPV), negative predictive value (NPV), detection rate (DR), misclassification error rate (MER), area under the ROC curve (AUC), and runtime.

Scientific Applications:

• Omics Data Preprocessing: Improves data quality in transcriptomics and metabolomics by robustly imputing missing values and reducing outlier influence prior to downstream analysis.

Methodology:

rMisbeta employs an iterative minimum beta divergence optimization procedure with robust mean and variance estimation to jointly address missing data and outliers, enhancing stability and accuracy in high-dimensional biological datasets.