def scoreClassifications(self):
total = 0
scores = []
for uid, actual in self.hand_classified_set.iteritems():
guess = self.algo_classified_set[uid]
scores.append(util.score_prediction(guess, actual))
total = sum(scores)
max_potential = len(self.hand_classified_set.keys()) * 6.0
unique, counts = np.unique(scores, return_counts=True)
frequencies = np.asarray((unique, counts)).T
return total, max_potential, frequencies
def scoreClassifications(self):
total = 0
scores = []
for uid, actual in self.hand_classified_set.iteritems():
guess = self.algo_classified_set[uid]
scores.append(util.score_prediction(guess, actual))
total = sum(scores)
max_potential = len(self.hand_classified_set.keys()) * 6.0
unique, counts = np.unique(scores, return_counts=True)
frequencies = np.asarray((unique, counts)).T
return total, max_potential, frequencies
def getPredictionScoreOfTrainingSet():
"""
Returns the score of the classifications on the validation set.
"""
hand_classified_set = loader.get_hand_classifiedset()
algo_classified_set = loader.get_algo_classifiedset()
total = 0
scores = []
for uid, actual in hand_classified_set.iteritems():
guess = algo_classified_set[uid]
scores.append(util.score_prediction(guess, actual))
total = sum(scores)
max_potential = len(hand_classified_set.keys()) * 6
return total / float(max_potential)
def getPredictionScoreOfTrainingSet():
"""
Returns the score of the classifications on the validation set.
"""
hand_classified_set = loader.get_hand_classifiedset()
algo_classified_set = loader.get_algo_classifiedset()
total = 0
scores = []
for uid, actual in hand_classified_set.iteritems():
guess = algo_classified_set[uid]
scores.append(util.score_prediction(guess, actual))
total = sum(scores)
max_potential = len(hand_classified_set.keys()) * 6
return total / float(max_potential)
test = pd.read_csv(pred_file, dtype=np.int32)
test['pred'] = test['ad_id'].map(ad_pos_cnt).fillna(0) / (
test['ad_id'].map(ad_cnt).fillna(0) + reg)
return test
## Validating
print "Running on validation split..."
pred = fit_predict(val_split[0], val_split[1])
print "Scoring..."
present_score, future_score, total_score = score_prediction(pred)
name = gen_prediction_name('ad-mean', total_score)
print " Present score: %.5f" % present_score
print " Future score: %.5f" % future_score
print " Total score: %.5f" % total_score
pred[['display_id', 'ad_id', 'pred']].to_pickle('preds/%s-val.pickle' % name)
## Predicting
print "Running on full split..."
pred = fit_predict(full_split[0], full_split[1])
pred[['display_id', 'ad_id', 'pred']].to_pickle('preds/%s-test.pickle' % name)
