def optimize_sgdc():
df, xnames, yname = load_df_from_pkl()
baseline_sgdc(df, xnames, yname)
t0 = time.time()
params_grid = [{"clf__n_iter": [5, 8, 10, 15],
"clf__alpha": [1e-7, 1e-6, 1e-5, 0.0001, 0.001, 0.01, 0.1, 1],
"clf__penalty": ["none", "l1", "l2"],
'clf__loss': ['log']}]
optmizer = ClassifierOptimizer('sgdc') # linear_model.SGDClassifier(loss='log')
optmizer.set_params(params_grid)
optmizer.add_pipleline([("scalar", StandardScaler())])
optmizer.optimize(df[xnames], df[yname], cv = 5, scoring = 'roc_auc')
print optmizer.get_score_gridsearchcv()
t1 = time.time() # time it
print "finish in %4.4fmin for %s " %((t1-t0)/60,'optimizer')
def optimize_rf(df, xnames, yname):
optmizer = ClassifierOptimizer('rf')
params = [{"clf__n_estimators": [100, 300]
, 'clf__max_depth':[5, 10, 20, 30]
, 'clf__min_samples_leaf':[5]
, 'clf__class_weight':[None, {0:1, 1:1}, {0:1,1:5}, {0:1,1:10}, 'auto'] }]
optmizer.set_params(params)
optmizer.optimize(df[xnames], df[yname], cv = 5, scoring = 'roc_auc')
print optmizer.get_score_gridsearchcv()