












Mahalanobis Distance
distance Given a positivedefinite symmetric matrix ¢² (a covariance
matrix) , the distance between examples x and y in feature space
is









MAP Estimator
is the global maximum of the posterior density. MAP stands for
maximum a posteriori.









MLE(Maximum Likelihood Estimatez)
is a value which maximum the likelihood function, or, more loosely,
is a stationary point of the likelihood function. 








Missing Values
are unreported values of features. These could be lost, or they
could be deliberate nonmeasurement and so convey information.
(A medical practitioner will not order tests if their value
appears low.) 








Multilayer Perceptron
is another name for a feedforward neural network. Despite the
name, the 'neurons' are not usually perceptrons. 








Multivariate Analysis
is the branch of statistics concerned with multiple observations
on each example which are not only of interest to predict or
explain just one of the observations. 








