Sparse Multinomial Logistic Regression: Fast Algorithms and Generalization Bounds
IEEE Transactions on Pattern Analysis and Machine Intelligence
IEEE Transactions on Information Technology in Biomedicine - Special section on biomedical informatics
IEEE Transactions on Fuzzy Systems
Hi-index | 0.00 |
In this paper, we applied a regularized multinomial logistic regression (RMLR) for multicolor fluorescence in-situ hybridization (M-FISH) image analysis, in order to better classify chromosomes. The RMLR integrates complementary information from multi-channel M-FISH images and considers the relationship of these data between different channels. We compared the model with two other regression models, e.g., multinomial logistic regression (MLR) and sparse multinomial logistic regression (SMLR). We show that the correct classification ratio of chromosomal region by the RMLR model is almost 93%, compared with 90% and 76% by the MLR and SMLR model when tested in a comprehensive M-FISH image database that we established and the p-value of these three models indicating that the RMLR model can significantly improve the accuracy of M-FISH image analysis.