T
T. Krishnan
Researcher at Indian Statistical Institute
Publications - 5
Citations - 56
T. Krishnan is an academic researcher from Indian Statistical Institute. The author has contributed to research in topics: Linear discriminant analysis & Normal distribution. The author has an hindex of 4, co-authored 5 publications receiving 56 citations.
Papers
More filters
Journal ArticleDOI
Efficiency of discriminant analysis when initial samples are classified stochastically
T. Krishnan,Subhas C. Nandy +1 more
TL;DR: The Efron efficiency of this procedure compared to the situation where the initial classification is done deterministically and correctly is studied, which concludes that stochastic supervision contains a great deal of information on the discriminant function.
Journal ArticleDOI
Efficiency of learning with imperfect supervision
TL;DR: Efron's Asymptotic Relative Efficiency is derived of the discriminant function estimated under this model, relative to the case when classification is perfect, and it is shown that training samples are useful even if prone to a certain amount of misclassification.
Journal ArticleDOI
Discriminant analysis with a stochastic supervisor
T. Krishnan,Subhas C. Nandy +1 more
TL;DR: The EM algorithm for maximum likelihood estimation of the parameters for linear discriminant analysis for two p -variate normal distributions with a common covariance matrix is derived.
Journal ArticleDOI
Pattern recognition with an imperfect supervisor
U. A. Katre,T. Krishnan +1 more
TL;DR: The likelihood equations for this case are derived and some ways of solving them to estimate parameters are discussed and results of computations with various choices of initial values of parameters in iterative procedures of estimation are reported.
Journal ArticleDOI
Efficiency of logistic-normal stochastic supervision
T. Krishnan,Subhas C. Nandy +1 more
TL;DR: The Efron efficiency of this stochastic supervision scheme, in which the logarithm of odds ratio of the supervisor classification has a normal distribution and is independent of the feature vector, is computed.