def interaction(self):
# mu
mu = np.zeros((self.mode_num, self.mode_num))
for i in range(self.mode_num):
for j in range(self.mode_num):
pij = (1 -
(self.mode_num - 1) * self.f) if (i == j) else self.f
mu[i, j] = pij * self.mode_dist[i]
c = np.sum(mu, 0)
mu = mu / c
pos = np.isnan(mu)
mu[pos] = 0
mode_particle_dict = {}
for j in range(self.mode_num):
particles = []
for i in range(self.mode_num):
for p in self.mode_particle_dict[j]:
_p = p.clone()
_p.set_weigth(_p.weight * mu[i, j])
particles.append(_p)
mode_particle_dict[j] = resample(particles, self.N)
self.mode_particle_dict = mode_particle_dict
return c
def pf_identify_fault_para(self, window=2.0):
i = self.detect_para_fault()
self.para_fault_id.append(i)
window_len = int(window/self.hsw.step_len)
if i!=0 and (np.array(self.para_fault_id[-window_len:])!=0).all() and \
abs(self.t-self.latest_sp)>3 and not self.particle_para_estimation:
self.particle_para_estimation = True
i -= 1
self.N = self.Nmax
particles = self.tracjectory[-1]
self.detect_time = self.t # detect time
self.fault_time = self.t - window
self.fault_para_flag[i] = 1
particles = resample(particles, self.Nmax)
for p in particles:
magnitude = np.random.uniform(0.05, 0.5)
init_para = np.zeros(len(self.hsw.para_faults()))
init_para[i] = magnitude
p.set_para(init_para)
self.tracjectory[-1] = particles
if not self.particle_para_estimation:
fault_para = self.fault_para
else:
particles = self.tracjectory[-1]
fault_para = np.sum([p.weight*p.para for p in particles], 0)
self.para.append(fault_para)
def inverse_transform(self, X):
"""
Inverse transform of sampled data X. It requires original_shape to be
set (via set_params).
"""
if self.original_rows is None:
raise ValueError("original shape is not set!")
newshape = (self.original_rows, X.shape[1])
return resample(X, newshape, self.method, self.centre, self.minusone)
def transform(self, X, y=None):
"""
Resample the matrix X.
X is of shape [n_samples, n_features]. It has to have at least 2
dimensions.
X can be a list of matrices, in which case tranforms is applied to each
of them independently and it returns a list of discretized matrices.
"""
if type(X) is list:
ret = []
for x in X:
newshape = (self.newshape, x.shape[1])
ret.append(resample(x, newshape, self.method,
self.centre, self.minusone)
)
return ret
else:
newshape = (self.newshape, X.shape[1])
return resample(X, newshape, self.method, self.centre, self.minusone)
def step(self, particle, obs):
self.t += self.hsw.step_len
particle_ip1 = []
for ptc in particle:
p = self.step_particle(ptc, obs)
particle_ip1.append(p)
normalize(particle_ip1, 0)
re_particle_ip1 = resample(particle_ip1, self.N)
ave_state = sum([p.weight * p.state for p in re_particle_ip1])
max_mode = [p.mode for p in re_particle_ip1]
num_counter = Counter(max_mode)
max_mode = num_counter.most_common(1)[0][0]
self.tracjectory.append(re_particle_ip1)
self.state.append(ave_state)
self.mode.append(max_mode)
def step(self, obs):
self.t += self.hsw.step_len
mode_i0 = self.mode0 if not self.state else self.mode[len(
self.state) - 1] # lastest mode
particle_ip1 = {}
for m in range(self.mode_num):
particle_ip1[m] = []
particle = self.mode_particle_dict[m]
for ptc in particle:
p = self.step_particle(ptc, obs, mode_i0)
particle_ip1[m].append(p)
weight = [
sum([p.weight for p in particle_ip1[m]]) for m in particle_ip1
]
weight = [w / sum(weight) for w in weight]
w_max = max(weight) # maximal weight
m_opt = weight.index(w_max) # optimal mode
new_particle_ip1 = {}
for m in range(self.mode_num):
if weight[m] == 0 or (w_max / weight[m] > self.r):
copy_ptc = []
for p in particle_ip1[m_opt]:
_p = p.clone()
_p.mode = m
copy_ptc.append(_p)
new_particle_ip1[m] = copy_ptc
else:
new_particle_ip1[m] = particle_ip1[m]
m_opt = m_opt if weight[mode_i0] == 0 or (
weight[m_opt] / weight[mode_i0] > self.r) else mode_i0
for m in range(self.mode_num):
normalize(new_particle_ip1[m], 0.0001)
new_particle_ip1[m] = resample(new_particle_ip1[m])
ave_state = sum([p.weight * p.state for p in new_particle_ip1[m_opt]])
self.tracjectory.append(new_particle_ip1[m_opt])
self.state.append(ave_state)
self.mode.append(m_opt)
self.mode_particle_dict = new_particle_ip1
def step(self, particles, obs, mode):
'''
particles: particle list
'''
self.t += self.hsw.step_len
obs_conf = self.obs_conf if (self.fault_para_flag==0).all() else 0.0
mode_i0 = self.mode0 if not self.state else self.mode[len(self.state)-1]
self.latest_sp = self.latest_sp if mode_i0==mode else self.t
particles_ip1 = []
res = np.zeros(len(self.hsw.obs_sigma))
for ptc in particles:
p, r = self.step_particle(ptc, obs, mode_i0, mode)
particles_ip1.append(p)
res += r
self.last_likelihood = sum([ptc.weight for ptc in particles_ip1])
normalize(particles_ip1, obs_conf)
re_particles_ip1 = resample(particles_ip1, self.N)
ave_state = self.ave_state(re_particles_ip1)
self.tracjectory.append(re_particles_ip1)
self.state.append(ave_state)
self.res.append(res)
self.process_fault(res)
