def MIMO_probability_estimate(self):
""" Simulates 1000 system responses with specified quantity of
noise, predicts response with saved model and plots prediction
and validation data points and coefficient of determination."""
if not(path.exists(self.errorCSV)):
print("No simulation for these parameters exists in Uncertainty data. Proceeding with simulation")
# Initialize the models that are saved using the parameters declared above
predictor = Model(self.nstep)
predictor.load_MIMO()
deviations = np.arange(0,self.maxError)
stdev = np.array([0])
error=np.array([0])
b_predicted = np.array([0])
a_predicted = np.array([0])
b_true_value = np.array([0])
a_true_value = np.array([0])
for deviation in deviations:
numTrials = self.numTrials; nstep = self.nstep
timelength = self.timelength; trainFrac = self.trainFrac
# then simulates using the initialized model
sig = Signal(numTrials,nstep,timelength,trainFrac,stdev=deviation)
sig.MIMO_simulation(b_possible_values=[1,10],a_possible_values=[1,10])
# In this case, since we are only loading the model, not trying to train it,
# we can use function simulate and preprocess
xData,yData = sig.MIMO_validation()
# Function to make predictions based off the simulation
predictor.predict_MIMO(sig,savePredict=False,plotPredict=False,probabilityCall=True)
error = np.concatenate((predictor.errors,error))
b_predicted = np.concatenate((predictor.b_model_predictionictions.ravel(),b_predicted))
a_predicted = np.concatenate((predictor.a_model_predictionictions.ravel(),a_predicted))
b_true_value = np.concatenate((sig.aVals.ravel(),b_true_value))
a_true_value = np.concatenate((sig.aVals.ravel(),a_true_value))
stdev = np.concatenate((np.full_like(predictor.errors,deviation),stdev))
# Transfer to pandas to save to CSV
sd = pd.DataFrame()
sd['b_model_prediction'] = b_predicted
sd['a_model_prediction'] = a_predicted
sd['b_true_value'] = b_true_value
sd['a_true_value'] = a_true_value
sd.to_csv(self.errorCSV, index=False)
# If the simulation does exist for this set of parameters,
# load from CSV and plot
else:
print("Data exists for the parameters, proceeding to producing uncertainty estimate")
try:
sd = pd.read_csv(self.errorCSV).drop(['Unnamed: 0'],axis=1)
sd.drop(sd.tail(1).index,inplace=True)
except:
sd = pd.read_csv(self.errorCSV)
sd.drop(sd.tail(1).index,inplace=True)
# Plot histogram and predicted/true parameters
self.errorDict = {}
pathPrefix = self.prefix+"MIMO/Plots/"
prefixes = ['b','a']
for (i,prefix) in enumerate(prefixes):
sd[prefix+'Error'] = (sd[prefix+'Pred']-sd[prefix+'True'])
h = np.std(sd[prefix+'Error'])
self.errorDict[prefix] = h
if self.plotUncertainty:
plt.figure(dpi=200)
plt.hist(sd[prefix+'Error'],bins=100,label='Max Error = %i%%' % self.maxError)
plt.xlabel('Standard Error in '+ prefix)
plt.ylabel("Frequency Distribution")
plt.legend()
savePath = pathPrefix + "histogram_" + prefix + ".png"
plt.savefig(savePath)
plt.figure(dpi=200)
plt.plot(sd[prefix+'True'],sd[prefix+'Pred'],'.',
label='Max Error = %i%%' % self.maxError)
plt.plot(np.linspace(1,10),np.linspace(1,10),'r--',
label="r\u00b2 = %.3f" % r2_score(sd[prefix+'True'],sd[prefix+'Pred']))
plt.plot(np.linspace(1,10),np.linspace(1,10)+h,'g--',label="Stdev = %.3f" % h)
plt.plot(np.linspace(1,10),np.linspace(1,10)-h,'g--')
plt.ylabel("Predicted Value of "+prefix)
plt.xlabel("True Value of "+prefix)
plt.legend()
savePath = pathPrefix + "determination_" + prefix + ".png"
plt.savefig(savePath)
