Linear discriminant analysis
Citation, DOI, disclosures and article data
At the time the article was created Candace Makeda Moore had no recorded disclosures.View Candace Makeda Moore's current disclosures
At the time the article was last revised Daniel J Bell had no recorded disclosures.View Daniel J Bell's current disclosures
Linear discriminant analysis (LDA) is a type of algorithmic model employed in machine learning in order to classify data. Unlike some other now popular models, linear discriminant analysis has been used for decades in both AI for radiology 1 and many other biomedical applications.
Linear discriminant analysis has several variations including regularized discriminant analysis and flexible discriminant analysis. If logistic regression can be conceived as a line that separates data into two categories based on two features, then the basic design of linear discriminant analysis can be understood as a curved hyperplane that separates data based on many features. Linear discriminant analysis assumes instances of data are understood as belonging to several categories, whether they be pixels in an image or images in a dataset, and have several features, and then finds the best planes along which to separate them. Although linear discriminant analysis is not appropriate for every type of data set, it is often used in many types of problems including segmentation, texture analysis, and the classification of different types of tumors or lesions 2,3.
- 1. Chan HP, Wei D, Helvie MA, Sahiner B, Adler DD, Goodsitt MM, Petrick N. Computer-aided classification of mammographic masses and normal tissue: linear discriminant analysis in texture feature space. (1995) Physics in medicine and biology. 40 (5): 857-76. doi:10.1088/0031-9155/40/5/010 - Pubmed
- 2. Shannon C. Agner, Mark A. Rosen, Sarah Englander, John E. Tomaszewski, Michael D. Feldman, Paul Zhang, Carolyn Mies, Mitchell D. Schnall, Anant Madabhushi. Computerized Image Analysis for Identifying Triple-Negative Breast Cancers and Differentiating Them from Other Molecular Subtypes of Breast Cancer on Dynamic Contrast-enhanced MR Images: A Feasibility Study. (2014) Radiology. 272 (1): 91-9. doi:10.1148/radiol.14121031 - Pubmed
- 3. Eduardo Fleury, Karem Marcomini. Performance of machine learning software to classify breast lesions using BI-RADS radiomic features on ultrasound images. (2019) European Radiology Experimental. 3 (1): 1. doi:10.1186/s41747-019-0112-7 - Pubmed
- 4. Li Y, Yan C, Weng S, Shi Z, Sun H, Chen J, Xu X, Ye R, Hong J. Texture analysis of multi-phase MRI images to detect expression of Ki67 in hepatocellular carcinoma. (2019) Clinical radiology. 74 (10): 813.e19-813.e27. doi:10.1016/j.crad.2019.06.024 - Pubmed