random.partition.kernel: Random Partition Kernel
In mattdneal/gaussianProcess: Gaussian Process Fitting and Prediction

Takes a partition function and returns the covariance between two points derived from that partition function. The partition function should return a vector of integers, with each integer representing the result of a different classifier.

1	random.partition.kernel(a, b, hyper.params = NULL, additional.params)

`a`	first data point
`b`	second data point
`hyper.params`	an empy vector (included for compatibility with kernel function format)
`additional.params`	A list with a named element "partitionFunction" containing the partition function which defines the kernel