def murphy(self):
tally = np.zeros(4)
counts = np.array([8, 8, 8, 8], float)
results = np.zeros((len(self.seeds), 4))
for seed in self.seeds:
stimuli = stimgen.Murphy(self.dim, self.dvals)
weight = weights.murphy(seed)
for stimulus in stimuli.stimuli:
print 'Testing ', stimulus
c = self.count(stimulus)
w = weight[0][c,:]
p = weight[1][c,:]
print 'Weights', w
print 'Products', p
if 'Prototype' in stimulus:
s = [1,1,1,1]
model = ConceptModel.murphy(self.dim,
s, stimuli,
w, c, seed,
raster=self.raster)
model.run()
elif 'ConsistentA' in stimulus:
s = [1,0,1,1]
model = ConceptModel.murphy(self.dim,
s, stimuli,
w, c, seed,
raster=self.raster)
model.run()
elif 'ConsistentB' in stimulus:
s = [0,1,1,1]
model = ConceptModel.murphy(self.dim,
s, stimuli,
w, c, seed,
raster=self.raster)
model.run()
else:
s = [0,0,0,1]
model = ConceptModel.murphy(self.dim,
s, stimuli,
w, c, seed,
raster=self.raster)
model.run()
if model.result == True:
label = self.murphy_labels(stimulus)
tally[label] += 1
print 'Status: ', np.where(self.seeds == seed)[0][0], \
stimulus, tally, seed
error = np.divide(tally, counts)
results[np.where(self.seeds == seed)[0][0], :] = error
tally = np.zeros(4)
average = np.divide(np.sum(results, axis=0), float(len(self.seeds)))
np.save('results/Murphy', average)
print average
return average
def posner(self):
tally = np.zeros(4)
total = np.zeros((len(self.seeds), 4))
counts = np.array([6, 3, 6, 6], float)
results = np.zeros((len(self.dvals), 4))
for dval in self.dvals:
for seed in self.seeds:
np.random.seed(seed)
stimuli = stimgen.Posner(self.dim, dval)
for stimulus in stimuli.test_vectors:
print 'Testing ', stimulus
model = ConceptModel.posner(self.dim,
stimulus,
stimuli,
seed,
raster=self.raster)
model.run()
if model.result == True:
label = self.posner_labels(stimulus)
tally[label] += 1
print 'Status: ', np.where(self.seeds == seed)[0][0], \
stimulus, tally, seed
error = 1 - np.divide(tally, counts)
total[np.where(self.seeds == seed)[0][0], :] += error
tally = np.zeros(4)
average = np.divide(np.sum(total, axis=0), float(len(self.seeds)))
results[np.where(self.dvals == dval)[0][0], :] = average
total = np.zeros((len(self.seeds), 4))
np.save('results/Posner' + str(self.dvals[0]), results)
print results
return results
def posner(self):
tally = np.zeros(4)
total = np.zeros((len(self.seeds), 4))
counts = np.array([6,3,6,6], float)
results = np.zeros((len(self.dvals), 4))
for dval in self.dvals:
for seed in self.seeds:
np.random.seed(seed)
stimuli = stimgen.Posner(self.dim, dval)
for stimulus in stimuli.test_vectors:
print 'Testing ', stimulus
model = ConceptModel.posner(self.dim,
stimulus,
stimuli,
seed,
raster=self.raster)
model.run()
if model.result == True:
label = self.posner_labels(stimulus)
tally[label] += 1
print 'Status: ', np.where(self.seeds == seed)[0][0], \
stimulus, tally, seed
error = 1 - np.divide(tally, counts)
total[np.where(self.seeds == seed)[0][0], :] += error
tally = np.zeros(4)
average = np.divide(np.sum(total, axis=0), float(len(self.seeds)))
results[np.where(self.dvals == dval)[0][0], :] = average
total = np.zeros((len(self.seeds), 4))
np.save('results/Posner'+str(self.dvals[0]), results)
print results
return results
def brooks(self):
total = np.zeros((len(self.seeds), 3))
counts = np.array([8, 4, 4], float)
tally = np.zeros(3)
results = np.zeros((len(self.dvals), 3))
for dval in self.dvals:
for seed in self.seeds:
np.random.seed(seed)
stimuli = stimgen.Brooks(self.dim, dval)
if np.where(self.seeds == seed)[0][0] < 4:
stimuli.rule1()
elif np.where(self.seeds == seed)[0][0] < 8:
stimuli.rule2()
elif np.where(self.seeds == seed)[0][0] < 12:
stimuli.rule3()
else:
stimuli.rule4()
for stimulus in stimuli.test_vectors:
print 'Testing ', stimulus
model = ConceptModel.brooks(self.dim,
stimulus,
stimuli,
seed,
raster=self.raster)
model.run()
if model.result == True:
label = self.brooks_labels(stimulus, stimuli)
tally[label] += 1
print 'Status: ', np.where(self.seeds == seed)[0][0], \
stimulus, tally, seed
error = 1 - np.divide(tally, counts)
total[np.where(self.seeds == seed)[0][0], :] = error
tally = np.zeros(3)
average = np.divide(np.sum(total, axis=0), float(len(self.seeds)))
results[np.where(self.dvals == dval)[0][0], :] = average
total = np.zeros((len(self.seeds), 3))
np.save('results/Brooks' + str(self.dvals[0]), results)
print results
return results
def brooks(self):
total = np.zeros((len(self.seeds), 3))
counts = np.array([8, 4, 4], float)
tally = np.zeros(3)
results = np.zeros((len(self.dvals), 3))
for dval in self.dvals:
for seed in self.seeds:
np.random.seed(seed)
stimuli = stimgen.Brooks(self.dim, dval)
if np.where(self.seeds == seed)[0][0] < 4:
stimuli.rule1()
elif np.where(self.seeds == seed)[0][0] < 8:
stimuli.rule2()
elif np.where(self.seeds == seed)[0][0] < 12:
stimuli.rule3()
else:
stimuli.rule4()
for stimulus in stimuli.test_vectors:
print 'Testing ', stimulus
model = ConceptModel.brooks(self.dim,
stimulus,
stimuli,
seed,
raster=self.raster)
model.run()
if model.result == True:
label = self.brooks_labels(stimulus, stimuli)
tally[label] += 1
print 'Status: ', np.where(self.seeds == seed)[0][0], \
stimulus, tally, seed
error = 1 - np.divide(tally, counts)
total[np.where(self.seeds == seed)[0][0], :] = error
tally = np.zeros(3)
average = np.divide(np.sum(total, axis=0), float(len(self.seeds)))
results[np.where(self.dvals == dval)[0][0], :] = average
total = np.zeros((len(self.seeds), 3))
np.save('results/Brooks'+str(self.dvals[0]), results)
print results
return results
def murphy(self):
tally = np.zeros(4)
counts = np.array([8, 8, 8, 8], float)
results = np.zeros((len(self.seeds), 4))
for seed in self.seeds:
stimuli = stimgen.Murphy(self.dim, self.dvals)
weight = weights.murphy(seed)
for stimulus in stimuli.stimuli:
print 'Testing ', stimulus
c = self.count(stimulus)
w = weight[0][c, :]
p = weight[1][c, :]
print 'Weights', w
print 'Products', p
if 'Prototype' in stimulus:
s = [1, 1, 1, 1]
model = ConceptModel.murphy(self.dim,
s,
stimuli,
w,
c,
seed,
raster=self.raster)
model.run()
elif 'ConsistentA' in stimulus:
s = [1, 0, 1, 1]
model = ConceptModel.murphy(self.dim,
s,
stimuli,
w,
c,
seed,
raster=self.raster)
model.run()
elif 'ConsistentB' in stimulus:
s = [0, 1, 1, 1]
model = ConceptModel.murphy(self.dim,
s,
stimuli,
w,
c,
seed,
raster=self.raster)
model.run()
else:
s = [0, 0, 0, 1]
model = ConceptModel.murphy(self.dim,
s,
stimuli,
w,
c,
seed,
raster=self.raster)
model.run()
if model.result == True:
label = self.murphy_labels(stimulus)
tally[label] += 1
print 'Status: ', np.where(self.seeds == seed)[0][0], \
stimulus, tally, seed
error = np.divide(tally, counts)
results[np.where(self.seeds == seed)[0][0], :] = error
tally = np.zeros(4)
average = np.divide(np.sum(results, axis=0), float(len(self.seeds)))
np.save('results/Murphy', average)
print average
return average