for testing and deploying your application
for finding and fixing issues
for empowering human code reviews
# Author: Simon Blanke
# Email: [email protected]
# License: MIT License
import os
import numpy as np
from sklearn.externals import joblib
def find_best_hyperpara(features, scores):
N_best_features = 1
scores = np.array(scores)
index_best_scores = list(scores.argsort()[-N_best_features:][::-1])
best_score = scores[index_best_scores][0]
best_features = features.iloc[index_best_scores]
return best_features, best_score
class Predictor:
def __init__(self):
pass
def search(self, X_test):
best_para, best_score = self._predict(X_test)
return best_para, best_score
def load_model(self, path):
if os.path.isfile(path):
self.meta_reg = joblib.load(path)
else:
print("File at path", path, "not found")
def _predict(self, X_test):
score_pred = self.meta_reg.predict(X_test)
best_features, best_score = find_best_hyperpara(X_test, score_pred)
SimonBlanke /
Hyperactive
