mabwiser 2.2.0

• Modified predict_expectations, such that predict can use predict_expectations directly in all non-contextual learning policies.

mabwiser 2.1.0

• Added warm_start method to MAB, that allows untrained (cold) arms to be warm started based on features of each arm.
• Added remove_arm method to MAB to allow arms to be removed from bandit.

mabwiser 2.0.0

• Breaking: Updated NumPy RNG backend to utilize the new Generator class. This is a breaking change for algorithms with random components.
• Updated NumPy version dependency to >=1.17.0 to reflect the utilization of the new Generator class.
• Updated multivariate sampling logic in LinTS to utilize updated NumPy RNG backend

mabwiser 1.12.0

• New Algorithm: TreeBandit as a decision-tree based neighborhood policy.

mabwiser 1.11.0

• New Algorithm: Locality Sensitive Hashing Approximate Nearest Neighbors

mabwiser 1.10.0

• Introduction of base random number generator class to abstract away RNG provider

mabwiser 1.9.0

• New Algorithm: Randomized Popularity (Popularity)

mabwiser 1.8.0

• New Algorithm: Linear Thompson Sampling (LinTS)

mabwiser 1.7.0

• The learning_policy and neighborhood_policy attributes have been changed to decorated properties.
• MAB objects can now be serialized with the pickle module.

mabwiser 1.6.0

• Configurable empty neighborhood operation for Radius policy
• Empty neighborhood operation changed to use numpy.random.choice instead of numpy.random.randint.