Ejemplo n.º 1
0
#     sorted_rocs = sorted(rocs, key=lambda x: x[1])
#     for tuple in sorted_rocs:
#         print tuple

    plt.hist(rocs)
    plt.title("Target AUC distribution")
    plt.xlabel("AUC")
    plt.ylabel("Frequency")
    plt.show()
    
    
parser = argparse.ArgumentParser()
parser.add_argument("filename", help="input file of scoredata")
file = parser.parse_args().filename
print file
data = pd.create_dict(file)

#TODO - take input file as commandline arg and update titles accordingly based on that
#data = pd.create_dict('SCOREDATA.vina.balanced')
#data = pd.create_dict('SCOREDATA.vina.reduced')
#data = pd.create_dict('SCOREDATA.dkoes.reduced')


#linreg_ccv_plot_roc(10)
#precision_recall_curve(10)
#rfc_test_on_train()
#bootstrap(10, 100)
#leave_target_out()
#ccv_plot_roc(10)

#logreg_precision_recall_ccv(10)
Ejemplo n.º 2
0
from joblib import delayed, Parallel
from scipy import interp
from sklearn.metrics import auc, roc_curve
from sklearn.linear_model import LinearRegression

__author__ = "Jesus Bracho"
__date__ = "01.29.2016"

# TODO: Old script. Refactor.


parser = argparse.ArgumentParser()
parser.add_argument("filename", help="input file of scoredata")
datafile = parser.parse_args().filename

data = pd.create_dict(datafile)
data.pop("fpps", None)

clf = LinearRegression(normalize=True)

# Initialize mean true positive and false positive rates.
mean_tpr = 0.0
mean_fpr = np.linspace(0, 1, 100)
all_tpr = []

# Retrievs a list of targets based on the data's key values.
targets = data.keys()
target_size = len(targets)
target_range = range(len(targets))

# Added to try and fix legend