class LMSAdaptiveFilter:
"""
The LMS Adaptive Filter.
"""
def __init__(self, order, damping=0.5):
self.order = order
self.damping = damping
self.X = Signal(order)
self.Y = Signal(order)
self.weights = [0] * order
def is_signal_outlier(self, sig):
X = np.array(self.X.signal)
weights = np.array(self.weights)
yest = weights.dot(X)
c = (1.0 * (sig - yest)) / (1. * X.dot(X))
weights = weights + self.damping * c * X
self.X.add(sig)
self.weights = list(weights)
return self._check_est(yest)
def _check_est(self, est):
if self.Y.can_use():
return est >= (2.0 * self.Y.sigma() + self.Y.mean())
return False
