Pattern Recognition, 4th Edition

Computer Programs

1. GMDAS algorithm. Write a MATLAB function named GMDAS that implements the GMDAS algorithm when normal distributions are adopted for the representation of the clusters. The function takes as input (a) an l × N dimensional matrix X whose columns are the data vectors, (b) an l × m dimensional matrix mv whose ith column contains an initial estimate of the mean of the ith normal distribution, (c) an l × l × m dimensional matrix mc, whose ith two-dimensional l × l component contains an initial estimate of the covariance matrix of the ith normal distribution, (d) a parameter e used in the termination condition of the algorithm, which is ‖Θ(t) − Θ(t − 1)‖ < e, (e) the maximum number of allowable iterations, maxiter, (f) a seed