-
Notifications
You must be signed in to change notification settings - Fork 0
/
nonLinearCompareAlgorithmResults.py
199 lines (154 loc) · 7.02 KB
/
nonLinearCompareAlgorithmResults.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
from sklearn import linear_model
from ExtractResult import Result
from Transformation import F1, F2, F3, F4, F5, F1_transf, F2_transf
from sklearn.grid_search import GridSearchCV
from LASSOModel import LASSOEstimator,Shooting
from sklearn.metrics import mean_squared_error
import numpy as np
from utility import get_current_data, assign_weights, compute_mse
import sys
folder = "AlgorithmResults/"
print(sys.argv[1])
sys.argv[1:] = [int(x) for x in sys.argv[1:]]
n_samples = sys.argv[1]
print("n_samples", n_samples)
original_features = sys.argv[2]
print("original_features", original_features)
transformation = F2()
print("function", transformation)
end = sys.argv[3]
print("lambda_max", end)
verbose = True
file = "nonLinearDataset/"+transformation.__class__.__name__+"/test"+transformation.__class__.__name__+"num_blocks_modified1000num_samples"+str(n_samples)+"n_features"+str(original_features)+"dynamic_set.npz"
results = Result(file, "lasso")
XTrain, YTrain, XTest, YTest,mses = results.extract_data()
#dict_ = results.extract_dict()
weights_data = results.extract_weights()
informative_indexes = results.extract_informative()
print (informative_indexes)
n_features = XTrain.shape[1]
index_mse = len(weights_data)-1
weights_data = weights_data[index_mse]
final_weights = np.zeros(original_features)
#keys_ = np.array(dict_.keys()).astype("int64")
#for key in keys_:
#final_weights[key] += np.sum(weights_data[dict_.get(key).astype("int64")])
ordered_final_weights = np.argsort(final_weights)[::-1]
if verbose:
print("-------------")
print("ranking of the featues:", ordered_final_weights)
print("-------------")
ordered_indexes = np.argsort(weights_data)[::-1]
if verbose:
print("position informarives", np.where(np.in1d(ordered_indexes, informative_indexes)==True)[0])
lasso_cv = linear_model.LassoCV(fit_intercept=False, max_iter=10000, n_jobs = -1)
lasso_cv.fit(XTrain,YTrain)
best_alpha = lasso_cv.alpha_
model = linear_model.Lasso(fit_intercept=False,alpha=best_alpha)
new_loss,beta,_ = compute_mse(model, XTrain, YTrain,XTest, YTest)
beta = np.abs(beta[:,0])
beta_ord = np.sort(beta)[::-1]
beta_ordered = beta_ord[beta_ord>=1.0]
len_div_zero = len(beta_ordered)
beta_indexes = np.argsort(np.abs(beta))[::-1][:len_div_zero]
real_indexes = []
# for b in beta_indexes:
# values_ = dict_.values()
# count=0
# for dict_value in values_:
# if b in dict_value:
# real_indexes.append(count)
# count+=1
if verbose:
print("real features:",real_indexes )
_, idx = np.unique(real_indexes, return_index=True)
sorted_idx = np.sort(idx)
if verbose:
print("unique real features", np.array(real_indexes)[sorted_idx])
print("valori dei beta ordinati", beta_ordered)
print("loss LASSO", new_loss)
print("------------------")
active_sets = np.arange(1,n_features+1)
#file1 = np.load("AlgorithmResults/Shooting_model_vary_active_set.npz")
#mses = file1["mses"]
#min_mse = np.argmin(mses)
#print(mses[min_mse])
#active_set = active_sets[min_mse]
#active_sets= []
#active_sets.append(active_set)
mses = []
beta_div_zeros = []
indexes_to_extract = []
ordered_indexes_div_zeros = []
current_informatives = []
alphas = np.linspace(0.0000001, end, 100)
parameters = {"alpha": alphas}
for active_set in active_sets:
ordered_values = np.sort(weights_data)[::-1][:active_set]
ordered_indexes = np.argsort(weights_data)[::-1][:active_set]
#active_indexes = extract_results(file)
weights = assign_weights(ordered_values.copy())
#weights[:70]= 0.0
#ordered_values = ordered_values[:len(weights)]
#ordered_indexes = ordered_indexes[:len(weights)]
#weights = assign_weights_ordered(ordered_values.copy())
#min_ = np.min(ordered_values)
#max_ = np.max(ordered_values)
#weights_std = (ordered_values - min_)/ (max_ - min_)
#weights = weights_std
#norm = Normalizer(norm = "max")
#weights_std = norm.fit_transform(ordered_values)[0]
#weights = 1-weights_std
#weights = weights[:index_mse*3]
current_informative = np.intersect1d(ordered_indexes, informative_indexes)
current_not_informative = np.array(list(set(ordered_indexes)-set(current_informative)))
if verbose:
print("informative", len(current_informative), "su", len(ordered_indexes))
print("non informative",len(current_not_informative),"su", len(ordered_indexes))
current_train, current_test = get_current_data(XTrain, XTest, ordered_indexes)
indexes_to_extract.append(ordered_indexes)
model_list = {Shooting(weights)}
ext_data = ".npz"
ext_model = ".pkl"
for model in model_list:
lasso = LASSOEstimator(model)
clf = GridSearchCV(lasso, parameters, fit_params = {"verbose" : False}, cv=3, scoring="mean_squared_error")
clf.fit(current_train, YTrain)
lambda_opt = clf.best_params_
if verbose:
print("best lambda", lambda_opt)
lasso.set_params(**lambda_opt)
lasso.fit(current_train,YTrain)
y_pred_train = lasso.predict(current_train)
mse_train = mean_squared_error(YTrain, y_pred_train)
if verbose:
print("mse_train "+model.__class__.__name__,mse_train)
y_pred_test = lasso.predict(current_test)
mse_test = mean_squared_error(YTest, y_pred_test)
if verbose:
print ("active_set", active_set, "mse_test "+model.__class__.__name__,mse_test)
mses.append(mse_test)
beta = lasso.beta
beta_zero = np.where(np.abs(beta)<0.1)[0]
ordered_indexes_zeros = ordered_indexes[beta_zero]
beta_div_zero = np.where(np.abs(beta)>=0.1)[0]
ordered_indexes_div_zero = ordered_indexes[beta_div_zero]
if verbose:
print(ordered_indexes_div_zero)
indexing = np.argsort(ordered_indexes_div_zero)
#print beta
beta_ordered_index = np.argsort(np.abs(beta))
#print(ordered_indexes[beta_ordered_index])
ordered_indexes_div_zeros.append(ordered_indexes_div_zero)
assert(len(beta_div_zero)+len(beta_zero)==len(beta))
inf_zero = np.intersect1d(ordered_indexes_zeros,current_informative)
non_inf_zero = np.intersect1d(ordered_indexes_zeros,current_not_informative)
if verbose:
print("beta inf a zero",len(inf_zero) , "su", len(current_informative))
print("beta non inf a zero", len(non_inf_zero), "su",len(current_not_informative))
print("----------------------------------------")
np.savez(folder+model.__class__.__name__+transformation.__class__.__name__+"model_features"+str(n_features)+"samples"+str(n_samples)+ext_data, informative_indexes = informative_indexes, ordered_indexes_div_zeros = ordered_indexes_div_zeros,XTrain = XTrain, XTest = XTest, YTrain = YTrain, YTest = YTest, indexes_to_extract = indexes_to_extract, beta_div_zeros = beta_div_zeros, active_sets = active_sets, mses = mses)
#joblib.dump(lasso, folder+model.__class__.__name__+'_model_2'+ext_model, compress=9)
min_mse = np.argmin(mses)
min_active_set = active_sets[min_mse]
print("loss minima", mses[min_mse], "con active_set", min_active_set)