A model-based background adjustment method for oligonucleotide expression arrays is introduced. The method improves the accuracy of gene expression measurements by accounting for non-specific hybridization and optical noise. The approach uses a statistical model to estimate background noise and adjusts probe intensities accordingly. The method is implemented in the Bioconductor project and has been shown to outperform existing methods in various applications. The model assumes that probe intensities consist of optical noise, non-specific hybridization, and specific hybridization. The model is fitted to data from experiments and publicly available datasets to estimate parameters. The method is applied to real data and shown to improve the accuracy of expression measurements. The software used to implement the method is available as part of the Bioconductor project. The method is compared to other approaches, including the default Affymetrix algorithm, and is shown to provide more accurate and precise results. The method is particularly effective for genes with low expression levels and for detecting differentially expressed genes. The model is based on a statistical framework that accounts for the variability in probe intensities and provides a robust adjustment for background noise. The method is described in detail, including the statistical model, estimation procedures, and practical applications. The results demonstrate that the method improves the performance of the technology in various practical applications.A model-based background adjustment method for oligonucleotide expression arrays is introduced. The method improves the accuracy of gene expression measurements by accounting for non-specific hybridization and optical noise. The approach uses a statistical model to estimate background noise and adjusts probe intensities accordingly. The method is implemented in the Bioconductor project and has been shown to outperform existing methods in various applications. The model assumes that probe intensities consist of optical noise, non-specific hybridization, and specific hybridization. The model is fitted to data from experiments and publicly available datasets to estimate parameters. The method is applied to real data and shown to improve the accuracy of expression measurements. The software used to implement the method is available as part of the Bioconductor project. The method is compared to other approaches, including the default Affymetrix algorithm, and is shown to provide more accurate and precise results. The method is particularly effective for genes with low expression levels and for detecting differentially expressed genes. The model is based on a statistical framework that accounts for the variability in probe intensities and provides a robust adjustment for background noise. The method is described in detail, including the statistical model, estimation procedures, and practical applications. The results demonstrate that the method improves the performance of the technology in various practical applications.