model_partials = [partial1_exponential, partial2_exponential]
model_pivotfunc = get_pivot_exponential
elif functionType == 'Logarithmic':
model_function = function_logarithmic
model_partials = [partial1_logarithmic]
model_pivotfunc = get_pivot_logarithmic
else:
#return
print("hi")
#Rejection tech determination:
rejTechInt = int(rejType)
r = rcr.RCR()
rejTech = ''
if symDist:
if equalCont:
r.setRejectionTech(rcr.SS_MEDIAN_DL)
rejTech = 'ss_med_dl'
elif oneSidedCont:
r.setRejectionTech(rcr.LS_MODE_68)
rejTech = 'ls_mode_68'
elif inBetweenCont:
r.setRejectionTech(rcr.LS_MODE_DL)
rejTech = 'ls_mode_dl'
else:
r.setRejectionTech(rcr.ES_MODE_DL)
rejTech = 'es_mode_dl'
ax.set_position([box.x0, box.y0, box.width * 0.65, box.height]) ax.legend(loc='center left', bbox_to_anchor=(1, 0.5)) plt.show() # get results pre-RCR contaminated_mu = np.mean(data) contaminated_sigma = np.std(data) print(contaminated_mu, contaminated_sigma) # perform RCR import rcr # initialize RCR with rejection technique: # (chosen from shape of uncontaminated + contaminated distribution) r = rcr.RCR(rcr.SS_MEDIAN_DL) r.performBulkRejection(data) # perform outlier rejection # View results post-RCR cleaned_mu = r.result.mu cleaned_sigma = r.result.stDev print(cleaned_mu, cleaned_sigma) # plot rejections cleaned_data = r.result.cleanY flags = r.result.flags # list of booleans corresponding to the original dataset, # true if the corresponding datapoint is not an outlier. cleaned_data_indices = r.result.indices
# for Travis import rcr data = [0, 0.1, -0.1, 0.2, -0.2, 1e50] r = rcr.RCR(rcr.LS_MODE_68) r.performBulkRejection(data) # perform outlier rejection # View results rejected_data = r.result.rejectedY assert len(rejected_data) >= 1 print(len(rejected_data))
return x - rcr.FunctionalForm.pivot
def get_pivot(xdata, weights, f, params):
sum = 0
for x in xdata:
sum += x
return sum / len(xdata)
# if __name__ == "__main__":
y = [0.1, 0.2, 0, 0.3, -0.1, -0.2, -0.3, 11]
w = [0.5, 1.0, 1.2, 3.0, 1.0, 0.6, 0.2, 1.0, 0.1, 0.05]
y = np.array(y)
w = np.array(w)
rcr_obj = rcr.RCR(rcr.ES_MODE_DL)
rcr_obj.performBulkRejection(w, y)
print(rcr_obj.result.flags)
tolerance = 0.01
guess = [2.1, 0.9]
y = [2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 20.0, 11.0]
x = [0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0]
w = [0.5, 1.0, 1.2, 3.0, 1.0, 0.6, 0.2, 1.0, 0.1, 2.0]
err_y = [0.1 for xx in x]
gaussianParams = [[float('nan'), float('nan')], [1.0, 2.0] ]
boundedParams = [[0.0, float('nan')], [float('nan'), float('nan')] ]
priors = rcr.Priors(rcr.MIXED_PRIORS, gaussianParams, boundedParams)
box = ax.get_position()
ax.set_position([box.x0, box.y0, box.width * 0.65, box.height])
ax.legend(loc='center left', bbox_to_anchor=(1, 0.5))
plt.show()
# initialize model and guess for model parameters
guess = [5, 1.5]
model = rcr.FunctionalForm(linear, xdata, ydata, [d_linear_1, d_linear_2],
guess)
# initialize and perform RCR
r = rcr.RCR(rcr.LS_MODE_68) # setting up for RCR with this rejection technique
r.setParametricModel(model) # tell RCR that we are model fitting
r.performBulkRejection(ydata) # perform RCR
# view results
best_fit_parameters = model.result.parameters # best fit parameters
rejected_data = r.result.rejectedY # rejected and non-rejected data
nonrejected_data = r.result.cleanY
nonrejected_indices = r.result.indices
print(best_fit_parameters)
# plot results