Bayesian_Optimization的使用（超参数选择）

安装BO软件包

1 pip install bayesian-optimization
2 conda install -c bayesian-optimization
3 git clone https://github.com/fmfn/BayesianOptimization.git

使用方法

#coding: utf-8

from bayes_opt import  BayesianOptimization
#建议、评估
#定义一个函数，边界并实例化一个对象
def black_box_function(x,y):
    return -x ** 2 - (y - 1) ** 2 + 1

optimizer = BayesianOptimization(
    f=None,
    pbounds={'x':(-2,2),'y':(-3,3)},
    verbose=2,
    random_state=1,)

from bayes_opt import UtilityFunction
utility = UtilityFunction(kind="ucb",kappa=2.5,xi=0.0)



next_point_to_probe = optimizer.suggest(utility)
print("Next point to probe is:", next_point_to_probe)

target = black_box_function(**next_point_to_probe)
print("Found the target value to be:", target)

optimizer.register(params=next_point_to_probe,target=target)



#最大化循环
for _ in range(5):
    next_point = optimizer.suggest(utility)
    target = black_box_function(**next_point)
    optimizer.register(params=next_point,target=target)
    print(next_point,target)
print(optimizer.max)

#处理离散参数
def func_with_discrete_params(x,y,d):
    #模拟 d 离散的必要
    assert type(d) == int
    return ((x+y+d) // (1+d) /(1+(x+y)**2))

def function_to_be_optimized(x,y,w):
    d = int(w)
    return func_with_discrete_params(x,y,d)

optimizer = BayesianOptimization(f=function_to_be_optimized,pbounds={'x': (-10, 10), 'y': (-10, 10), 'w': (0, 5)},verbose=2,random_state=1,)

optimizer.maximize(alpha=1e-3)


#调整基本高斯过程

optimizer = BayesianOptimization(
    f=black_box_function,
    pbounds={'x': (-2, 2), 'y': (-3, 3)},
    verbose=2,
    random_state=1,
)
optimizer.maximize(
    init_points=1,
    n_iter=5,
    # 下面为GP 回归参数
    alpha=1e-3,
    n_restarts_optimizer=5
)


optimizer.set_gp_params(normalize_y=True)

from bayes_opt.event import DEFAULT_EVENTS,Events
optimizer = BayesianOptimization(
    f=black_box_function,
    pbounds={'x': (-2, 2), 'y': (-3, 3)},
    verbose=2,
    random_state=1,
)


class BasicObserver:
    def update(self, event, instance):
        """Does whatever you want with the event and `BayesianOptimization` instance."""
        print("Event `{}` was observed".format(event))

my_observer = BasicObserver()

optimizer.subscribe(
    event=Events.OPTMIZATION_STEP,
    subscriber=my_observer,
    callback=None, # 此处使用 update 方法回调
)

def my_callback(event, instance):
    print("Go nuts here!")

optimizer.subscribe(
    event=Events.OPTMIZATION_START,
    subscriber="Any hashable object",
    callback=my_callback,
)
optimizer.maximize(init_points=1, n_iter=2)
#可以通过DEFAULT_EVENTS列出所有默认事件的列表
DEFAULT_EVENTS

更多使用方法请参考https://github.com/fmfn/BayesianOptimization