Mahalanobis’ residual-distance statistics, better known as Mahalanobis D^2 statistic measures the distance of a multi-variate (unknown sample set) from a known unit. It is different from Eclidean distance. [Further details given in attached document: 'Indian contribution to Statistics'].
1. Can u please update on the recent works/applications (including studies on its limitations, if any) of such "distance-measure"?
2. Also please profile the latest work/ applications of Rao-Cramer bound/ inequality. It is a pioneering work from an Indo-American living legend, Prof. CR Rao. (Pl. refer the attached profile).
The Mahalanobis' distance (pseudo metric) has found its way into quite a number of application areas. For example, in his M.Sc. thesis, Bahadir Durak introduces the application of this metric in classifying biomedical data (see attached pdf file).
Durak observes that data types and the relationships between them determine the type of distance measure (p. 15). For example, in biomedical data and the relations between them, there are correlations within the features. The biomedical data sets considered are numerical and Mahalanobis distance was chosen to take into account the correlations required for numerical medical data.
- Badges
- Science topic
- Similar topics
- Microbiology
- Scientific Research