diskriminantratio

Diskriminantratio is a term sometimes used in statistics and econometrics to describe the ratio of the determinant of the within-group scatter matrix to the determinant of the total scatter matrix. It is closely related to concepts like principal component analysis (PCA) and discriminant analysis. The diskriminantratio provides a measure of how well the data can be separated into different groups. A higher diskriminantratio generally indicates better separability between the groups.

This ratio is particularly relevant in the context of dimensionality reduction when the goal is to find a lower-dimensional representation that preserves or enhances group distinctions. It can be seen as an indicator of the informativeness of the data with respect to group membership. In some applications, the diskriminantratio can be used to select the optimal number of dimensions to retain in a dimensionality reduction technique. The calculation involves obtaining the within-group scatter matrix, which captures the variance within each group, and the total scatter matrix, which represents the overall variance of the data. The ratio of their determinants is then computed. While not a universally standard term, it appears in specific literature and methodologies focused on data analysis and classification.