class Chain:
Refer to the hierarchical Bayesian model discussed in the Bayesian-EUCLID paper (Fig. 2) for details regarding the parameters.
Attributes
-p0 - Numpy array of length chain_length containing different values of \(p_0\) in the chain
-vs - Numpy array of length chain_length containing different values of \(\nu_s\) in the chain
-sig2 - Numpy array of length chain_length containing different values of \(\sigma^2\) in the chain
-z - Numpy array of dimension chain_lengthXnumFeatures (see features_library.py) containing different values of z (activity) in the chain
-theta - Numpy array of dimension chain_lengthXnumFeatures (see features_library.py) containing different values of theta (feature coefficients) in the chain
-chain_length
-burn - Number of elements of the chain discarded as burn in in sampling the posterior probability distribution.
Methods
-__init__(...): - Generates an object of class Chain
-update_state(...): - Populates the chain with newly sampled state variables
-combine_chain(...): - Combines different parallelly generated Markov chains
-burn_chain(...): - Deletes the first burn number of elements of the chains