def train(self, features, y, pred=False):
# y:-1 or 1, features:[(index, value), (index, value), ...]
total_mean, total_variance = self._active_mean_variance(features)
v, w = util.gaussian_corrections(y * total_mean / np.sqrt(total_variance))
for feature_index, feature_value in features:
weight = self._get_weight(feature_index)
mean_delta = y * weight._variance / np.sqrt(total_variance) * v * feature_value
variance_multiplier = 1.0 - weight._variance / total_variance * w * (feature_value ** 2)
weight.update(mean_delta, variance_multiplier)
self._apply_dynamics(weight)
self._set_weight(feature_index, weight)
if pred:
return norm.cdf(total_mean / total_variance)
def train(self, features, label):
logger.info("Training: %s, %s features", label, len(features))
assert len(features) == self._config.num_features
y = util.label_to_float(label)
total_mean, total_variance = self._active_mean_variance(features)
v, w = util.gaussian_corrections(y * total_mean / np.sqrt(total_variance))
for feature in features:
weight = self._get_weight(feature)
mean_delta = y * weight.variance / np.sqrt(total_variance) * v
variance_multiplier = 1.0 - weight.variance / total_variance * w
updated = pb.Gaussian(
mean=weight.mean + mean_delta,
variance=weight.variance * variance_multiplier)
self._set_weight(feature, self._apply_dynamics(updated))
def train(self, features, label):
logger.info("Training: %s, %s features", label, len(features))
assert len(features) == self._config.num_features
y = util.label_to_float(label)
total_mean, total_variance = self._active_mean_variance(features)
v, w = util.gaussian_corrections(y * total_mean /
np.sqrt(total_variance))
for feature in features:
weight = self._get_weight(feature)
mean_delta = y * weight.variance / np.sqrt(total_variance) * v
variance_multiplier = 1.0 - weight.variance / total_variance * w
updated = pb.Gaussian(mean=weight.mean + mean_delta,
variance=weight.variance *
variance_multiplier)
self._set_weight(feature, self._apply_dynamics(updated))