Commit 3f926582 authored by Christian Marius Lillelund's avatar Christian Marius Lillelund
Browse files

updated compliance case, now success if >= 3.7

parent 381f00f5
Pipeline #107010 canceled with stage
in 15 minutes and 45 seconds
......@@ -379,7 +379,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.8.11"
"version": "3.8.8"
},
"orig_nbformat": 2
},
......
Method,c_harrell,c_uno,brier_score
GradientBoosting,0.7192826813205445,0.8093132425965777,0.14487422609218523
clf,version,accuracy_mean,accuracy_std,precision_mean,precision_std,recall_mean,recall_std,roc_auc_mean,roc_auc_std,pr_auc_mean,pr_auc_std,f1_mean,f1_std
KNN,Embedded,69.972,0.011,73.546,0.007,88.532,0.008,71.116,0.016,82.909,0.014,80.345,0.007
SVM,Embedded,69.494,0.003,69.575,0.002,99.514,0.002,70.459,0.011,84.206,0.008,81.894,0.002
LR,Embedded,71.152,0.002,71.582,0.003,96.84,0.007,72.068,0.01,85.501,0.007,82.314,0.001
XGB,Embedded,71.264,0.003,72.266,0.002,95.016,0.003,70.929,0.014,84.333,0.01,82.093,0.002
RF,Embedded,71.826,0.014,75.34,0.015,88.331,0.015,75.044,0.01,87.01,0.008,81.3,0.008
MLP,Embedded,70.843,0.007,71.641,0.004,95.908,0.008,72.228,0.009,85.674,0.005,82.016,0.004
clf,version,accuracy_mean,accuracy_std,precision_mean,precision_std,recall_mean,recall_std,roc_auc_mean,roc_auc_std,pr_auc_mean,pr_auc_std,f1_mean,f1_std
KNN,Embedded,77.835,0.02,79.774,0.019,85.686,0.021,84.251,0.016,88.037,0.018,82.606,0.016
SVM,Embedded,70.422,0.013,71.094,0.012,87.536,0.043,76.67,0.012,83.759,0.009,78.393,0.014
LR,Embedded,71.393,0.019,72.721,0.018,85.69,0.044,79.617,0.023,86.508,0.018,78.594,0.018
XGB,Embedded,70.989,0.01,72.605,0.018,85.028,0.037,77.139,0.02,84.093,0.021,78.246,0.009
RF,Embedded,80.469,0.018,83.044,0.019,85.753,0.014,86.524,0.021,89.897,0.015,84.366,0.014
MLP,Embedded,73.418,0.023,77.24,0.024,80.606,0.038,80.12,0.023,86.707,0.019,78.814,0.019
......@@ -23,7 +23,14 @@ def main():
target_settings).load_data()
X, y = dl.get_data()
model = GradientBoostingSurvivalAnalysis(n_estimators=200,
model = GradientBoostingSurvivalAnalysis(n_estimators=100,
learning_rate=0.1,
max_depth=14,
loss='coxph',
min_samples_split=4,
max_features='log2',
subsample=0.4,
dropout_rate=0.7,
random_state=0)
model.fit(X, y)
......
......@@ -21,14 +21,14 @@ def main():
params = {"n_estimators": 400,
"booster": "gbtree",
"max_depth": 5,
"gamma": 5,
"colsample_bytree": 1,
"min_child_weight": 8,
"reg_alpha": 10,
"max_depth": 8,
"gamma": 0,
"colsample_bytree": 0.3,
"min_child_weight": 9,
"reg_alpha": 18,
"reg_lambda": 0.9,
"learning_rate": 0.05,
"subsample": 0.8,
"subsample": 1,
"use_label_encoder": False,
"eval_metric": "logloss",
"objective": "binary:logistic",
......
......@@ -160,7 +160,7 @@ def accumulate_screenings(df: pd.DataFrame, settings: dict) -> pd.DataFrame:
number_session += 1
df.loc[row_week[0], 'HasCompletedSession'] = 1
df.loc[row_week[0], 'Baseline'] = 1
if row_mean_evaluation[1] > 4:
if row_mean_evaluation[1] >= 3.7:
df.loc[row_week[0], 'GotComplianceInSession'] = 1
else:
df.loc[row_week[0], 'GotComplianceInSession'] = 0
......
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment