willengler-uc
commited on
Commit
•
736c16d
1
Parent(s):
bfbadb9
Upload model_predict_df.py
Browse files- model_predict_df.py +33 -5
model_predict_df.py
CHANGED
@@ -39,8 +39,8 @@ def get_barrier(df_test):
|
|
39 |
|
40 |
return barriers
|
41 |
|
42 |
-
def
|
43 |
-
d = 'ASR_model/
|
44 |
scaler = joblib.load(os.path.join(d, 'StandardScaler.pkl'))
|
45 |
model = joblib.load(os.path.join(d, 'RandomForestRegressor.pkl'))
|
46 |
df_features = pd.read_csv(os.path.join(d, 'X_train.csv'))
|
@@ -65,6 +65,27 @@ def get_asr(df_test):
|
|
65 |
|
66 |
return asrs, asr_ebars
|
67 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
68 |
def process_data(comp_list, elec_list):
|
69 |
X = pd.DataFrame(np.empty((len(comp_list),)))
|
70 |
y = pd.DataFrame(np.empty((len(comp_list),)))
|
@@ -136,15 +157,22 @@ def make_predictions(comp_list, elec_list):
|
|
136 |
barriers = get_barrier(df_test)
|
137 |
|
138 |
df_test['ML pred ASR barrier (eV)'] = barriers
|
139 |
-
|
|
|
|
|
|
|
|
|
|
|
140 |
|
141 |
pred_dict = {'Compositions': comp_list,
|
142 |
'Electrolytes': elec_list,
|
143 |
'Cost ($/kg)': costs,
|
144 |
'Stability @ 500C (meV/atom)': stabilities,
|
145 |
'ASR barrier (eV)': barriers,
|
146 |
-
'log ASR at 500C (Ohm-cm2)': asrs,
|
147 |
-
'log ASR error (Ohm-cm2)': asr_ebars
|
|
|
|
|
148 |
|
149 |
return pd.DataFrame(pred_dict)
|
150 |
|
|
|
39 |
|
40 |
return barriers
|
41 |
|
42 |
+
def get_asr_rf(df_test):
|
43 |
+
d = 'ASR_model/ASR_RF_model'
|
44 |
scaler = joblib.load(os.path.join(d, 'StandardScaler.pkl'))
|
45 |
model = joblib.load(os.path.join(d, 'RandomForestRegressor.pkl'))
|
46 |
df_features = pd.read_csv(os.path.join(d, 'X_train.csv'))
|
|
|
65 |
|
66 |
return asrs, asr_ebars
|
67 |
|
68 |
+
def get_asr_gpr(df_test):
|
69 |
+
d = 'ASR_model/ASR_GPR_model'
|
70 |
+
scaler = joblib.load(os.path.join(d, 'StandardScaler.pkl'))
|
71 |
+
model = joblib.load(os.path.join(d, 'GaussianProcessRegressor.pkl'))
|
72 |
+
df_features = pd.read_csv(os.path.join(d, 'X_train.csv'))
|
73 |
+
|
74 |
+
features = df_features.columns.tolist()
|
75 |
+
df_test = df_test[features]
|
76 |
+
|
77 |
+
X_ASR = scaler.transform(df_test)
|
78 |
+
|
79 |
+
asrs, errs_list = model.model.predict(X_ASR, return_std=True)
|
80 |
+
|
81 |
+
# Get ebars and recalibrate them
|
82 |
+
a = 1.18033360971506
|
83 |
+
b = -0.0660773887574826
|
84 |
+
|
85 |
+
asr_ebars = a * np.array(errs_list) + b
|
86 |
+
|
87 |
+
return asrs, asr_ebars
|
88 |
+
|
89 |
def process_data(comp_list, elec_list):
|
90 |
X = pd.DataFrame(np.empty((len(comp_list),)))
|
91 |
y = pd.DataFrame(np.empty((len(comp_list),)))
|
|
|
157 |
barriers = get_barrier(df_test)
|
158 |
|
159 |
df_test['ML pred ASR barrier (eV)'] = barriers
|
160 |
+
|
161 |
+
# Get the ML (RF) predicted ASRs
|
162 |
+
asrs, asr_ebars = get_asr_rf(df_test)
|
163 |
+
|
164 |
+
# Get the ML (GPR) predicted ASRs
|
165 |
+
asrs_gpr, asr_ebars_gpr = get_asr_gpr(df_test)
|
166 |
|
167 |
pred_dict = {'Compositions': comp_list,
|
168 |
'Electrolytes': elec_list,
|
169 |
'Cost ($/kg)': costs,
|
170 |
'Stability @ 500C (meV/atom)': stabilities,
|
171 |
'ASR barrier (eV)': barriers,
|
172 |
+
'log ASR at 500C (Ohm-cm2) (RF)': asrs,
|
173 |
+
'log ASR error (Ohm-cm2) (RF)': asr_ebars,
|
174 |
+
'log ASR at 500C (Ohm-cm2) (GPR)': asrs_gpr,
|
175 |
+
'log ASR error (Ohm-cm2) (GPR)': asr_ebars_gpr}
|
176 |
|
177 |
return pd.DataFrame(pred_dict)
|
178 |
|