def get_param_grids():
lrs = [10**i for i in range(-4, -2)]
gd_params = {'learning_rate': [0.0001], 'max_iters': [2000]}
rhc_params = {
'learning_rate': [0.1],
'restarts': [100], #[10]
'max_iters': [2000]
}
sa_params = {
'learning_rate': [0.0001, 0.001, 0.01, 0.1],
'schedule': [
ExpDecay(),
ExpDecay(exp_const=0.002),
ExpDecay(exp_const=1 / 2000),
GeomDecay(),
GeomDecay(decay=.999),
ArithDecay(),
ArithDecay(decay=1 / 2000),
],
'max_iters': [2000]
}
ga_params = {
'learning_rate': [0.0001, 0.001, 0.01, 0.1],
'pop_size': [400], #[200, 400], #[200]
'mutation_prob': [0.1, .2, .5, .9],
'max_iters': [500]
}
params = {
'gradient_descent': gd_params,
'random_hill_climb': rhc_params,
'simulated_annealing': sa_params,
'genetic_alg': ga_params
}
return params
def test_exp_below_min():
"""Test exponential decay evaluation function for case where result is
below the minimum"""
schedule = ExpDecay(init_temp=10, exp_const=0.05, min_temp=1)
x = schedule.evaluate(50)
assert x == 1
def test_exp_above_min():
"""Test exponential decay evaluation function for case where result is
above the minimum"""
schedule = ExpDecay(init_temp=10, exp_const=0.05, min_temp=1)
x = schedule.evaluate(5)
assert round(x, 5) == 7.78801
def extract_best_hyperparameters(gridsearch_results):
"""extract_best_hyperparameters
Parameters
----------
gridsearch_results : dataframe of gridsearch results
Returns
the best hyperparameters as a dictionary
-------
"""
params = [col for col in gridsearch_results.columns if col not in METRICS]
best_row = gridsearch_results["Best Fitness"].idxmax()
best_params = gridsearch_results[params].iloc[best_row].to_dict()
for key, value in best_params.items():
if key == "schedule": # Special case of conversion
best_params[key] = ExpDecay(exp_const=value)
if value >= 1:
best_params[key] = int(
value
) # This might not be very safe but it works for these cases
return best_params
def get_param_grids():
lrs = [10**i for i in range(-4, -2)]
gd_params = {'learning_rate': lrs, 'max_iters': [2000]}
rhc_params = {
'learning_rate': lrs,
'restarts': [10, 30], #[10]
'max_iters': [2000]
}
sa_params = {
'learning_rate': lrs,
'schedule': [GeomDecay(), ArithDecay(),
ExpDecay()],
'max_iters': [2000]
}
ga_params = {
'learning_rate': lrs,
'pop_size': [200], #[200]
'mutation_prob': [0.05, 0.1, 0.2], #[0.1]
'max_iters': [500]
}
params = {
'gradient_descent': gd_params,
'random_hill_climb': rhc_params,
'simulated_annealing': sa_params,
'genetic_alg': ga_params
}
return params
def extract_best_hyperparameters(algo, output_dir):
if algo == "gs":
return best_nn_params
filename = os.path.join(output_dir, f"{algo}_nn_gsresults.csv")
df = pd.read_csv(filename, index_col=0)
params = [col for col in df.columns if col not in METRICS + ["Fold"]]
grouped = df.groupby(params)
mean_groups = grouped.mean().reset_index()
max_ix = mean_groups["Val_F1"].argmax()
best_params = mean_groups.loc[max_ix, params].to_dict()
for key, value in best_params.items():
if key == "schedule":
best_params[key] = ExpDecay(exp_const=value)
elif value.is_integer():
best_params[key] = int(value)
return best_params
def get_param_grids():
lrs = [10**i for i in range(-4, -2)]
gd_params = {'learning_rate': [0.0001], 'max_iters': [40]}
rhc_params = {
'learning_rate': [0.1],
'restarts': [100], #[10]
'max_iters': [4000]
}
sa_params = {
'learning_rate': [0.1],
'schedule': [
#AirithDecay(),
ExpDecay(),
#ExpDecay(exp_const=0.01),
#ExpDecay(exp_const=0.001),
#ExpDecay(exp_const=0.002),
#ExpDecay(exp_const=0.005),
#ExpDecay(exp_const=0.002, init_temp=.9),
#ExpDecay(exp_const=0.002, init_temp=.8),
#ExpDecay(exp_const=0.002, init_temp=.7),
],
'max_iters': [4000]
}
ga_params = {
'learning_rate': [0.1], #, 0.01],
'pop_size': [400], #[200, 400], #[200]
'mutation_prob': [0.9], #[0.1]
'clip_max': [100],
'max_iters': [1000]
}
params = {
'gradient_descent': gd_params,
'random_hill_climb': rhc_params,
'simulated_annealing': sa_params,
'genetic_alg': ga_params
}
return params
"Time Elapsed",
"Best Fitness",
"Fitness Evaluations",
"Global Optimum",
"Found Global Optimum",
]
OUTPUT_DIR = "output"
MAX_RESTARTS = 10
GRID = {
"rhc": {
"restarts": [10]
},
"sa": {
"schedule": [
ExpDecay(exp_const=0.001),
ExpDecay(exp_const=0.005),
ExpDecay(exp_const=0.01),
ExpDecay(exp_const=0.05),
ExpDecay(exp_const=0.1),
],
},
"ga": {
"pop_size": [200],
"mutation_prob": [0.05, 0.1, 0.2, 0.4]
},
"mimic": {
"pop_size": [200],
"keep_pct": [0.1, 0.3, 0.5, 0.7, 0.9]
},
}
def get_optimizers():
return {
'RHC': {
'function': random_hill_climb,
'kwargs': [{}],
},
'SA': {
'function':
simulated_annealing,
'kwargs': [
{
'schedule': ExpDecay()
},
{
'schedule': ExpDecay(exp_const=0.01)
},
{
'schedule': ExpDecay(exp_const=0.002)
},
{
'schedule': GeomDecay()
},
{
'schedule': GeomDecay(decay=.98)
},
{
'schedule': GeomDecay(decay=.96)
},
{
'schedule': ArithDecay()
},
{
'schedule': ArithDecay(decay=0.001)
},
{
'schedule': ArithDecay(decay=1 / 5000)
},
{
'schedule': ExpDecay(exp_const=0.002)
},
]
},
'GA': {
'function':
genetic_alg,
'kwargs': [
{
'pop_size': 200,
'mutation_prob': 0.1
},
{
'pop_size': 200,
'mutation_prob': 0.2
},
{
'pop_size': 200,
'mutation_prob': 0.05
},
{
'pop_size': 800,
'mutation_prob': 0.1
},
]
},
'MIMIC': {
'function':
mimic,
'kwargs': [
{
'pop_size': 200,
'keep_pct': 0.2,
#'fast_mimic': True
},
{
'pop_size': 200,
'keep_pct': 0.4,
#'fast_mimic': True
},
{
'pop_size': 400,
'keep_pct': 0.2,
},
]
}
}
def get_optimizers():
return {
#'RHC': {
# 'function': random_hill_climb,
# 'kwargs': [
# {}
# ],
#},
'SA': {
'function':
simulated_annealing,
'kwargs': [
{
'schedule': ExpDecay()
},
#{'schedule': ExpDecay(exp_const=0.01)},
{
'schedule': ExpDecay(exp_const=0.002)
},
{
'schedule': ExpDecay(exp_const=1 / 5000)
},
{
'schedule': GeomDecay()
},
#{'schedule': GeomDecay(decay=.98)},
{
'schedule': GeomDecay(decay=.9993)
},
{
'schedule': ArithDecay()
},
#{'schedule': ArithDecay(decay=0.001)},
{
'schedule': ArithDecay(decay=1 / 5000)
},
]
},
#'GA': {
# 'function': genetic_alg,
# 'kwargs': [
# #{'pop_size': 200,
# # 'mutation_prob': 0.1
# #},
# #{'pop_size': 200,
# # 'mutation_prob': 0.2
# #},
# #{'pop_size': 200,
# # 'mutation_prob': 0.05
# #},
# #{'pop_size': 400,
# # 'mutation_prob': 0.4
# #},
# #{'pop_size': 400,
# # 'mutation_prob': 0.2
# #},
# #{'pop_size': 400,
# # 'mutation_prob': 0.1
# #},
# {'pop_size': 800,
# 'mutation_prob': 0.1
# },
# ]
#},
#'MIMIC': {
# 'function': mimic,
# 'kwargs': [
# {'pop_size': 400,
# 'keep_pct': 0.2,
# },
# {'pop_size': 400,
# 'keep_pct': 0.1,
# },
# {'pop_size': 400,
# 'keep_pct': 0.05,
# },
# #{'pop_size': 800,
# # 'keep_pct': 0.2,
# #},
# #{'pop_size': 800,
# # 'keep_pct': 0.1,
# #},
# ]
#}
}