2019/11/7 Assignment 4
https://notebooks.dmaitre.phyip3.dur.ac.uk/miscada-da/user/czjs88/notebooks/Homework 4/Assignment 4.ipynb 1/10
In [ ]:
Question 1: Linear Regression (10 marks)
The data listed below come from an experiment to verify Ohm's law. The voltage across a resistor (the
dependent variable) was measured as a function of the current flowing (the independent variable). The
precision of the voltmeter was 0.01mV, and the uncertainty in the current was negligible.
This data is saved as 'ohms_law.csv'
Required:
(i) Calculate the unweighted best-fit gradient and intercept, and their uncertainties. (2 marks)
(ii) Calculate the common uncertainty, , and compare the value with the experimental
uncertainty. (2 marks)
(iii) Plot a graph of the data and add the best-fit straight line. Remember to label your axes. (4
marks)
(iv) Calculate the residuals, and comment on their magnitudes. (2 marks)
Current (μA)
Voltage (mV)
Current (μA)
Voltage (mV)
(i) Calculate the unweighted best-fit gradient and intercept, and their
uncertainties.
%matplotlib nbagg
from __future__ import division
from scipy.stats import norm
from IPython.display import HTML
from IPython.display import display
from scipy.optimize import *
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
2019/11/7 Assignment 4
https://notebooks.dmaitre.phyip3.dur.ac.uk/miscada-da/user/czjs88/notebooks/Homework 4/Assignment 4.ipynb 2/10
(ii) Calculate the common uncertainty (inside the function) and compare with the
experimental uncertainty (by typing your comparison in the markdown cell below
the function). You must return your answer in micro volts.
In [ ]:
YOUR ANSWER HERE
(iii) Plot a graph of the data adding the best fit line and residuals below.
data = pd.read_csv('ohms_law.csv')
current = np.array(data.iloc[:,0])
voltage = np.array(data.iloc[:,1])
voltage_error = data.iloc[:,2]
def linear(x,m,c):
return x*m + c
def one_i():
gradient = 0
intercept = 0
uncertainty_gradient = 0
uncertainty_intercept = 0
# YOUR CODE HERE
return(gradient,intercept,uncertainty_gradient,uncertainty_intercept)
print(one_i())
'''TEST CELL- DO NOT DELETE'''
'''TEST CELL- DO NOT DELETE'''
'''TEST CELL- DO NOT DELETE'''
'''TEST CELL- DO NOT DELETE'''
data = pd.read_csv('ohms_law.csv')
current = np.array(data.iloc[:,0])
voltage = np.array(data.iloc[:,1])
voltage_error = data.iloc[:,2]
def one_ii():
# YOUR CODE HERE
return(common_uncertainty)
print(one_ii())
2019/11/7 Assignment 4
https://notebooks.dmaitre.phyip3.dur.ac.uk/miscada-da/user/czjs88/notebooks/Homework 4/Assignment 4.ipynb 3/10
In [ ]:
(iv) Comment on the magnitude of the residuals.
YOUR ANSWER HERE
Question 2: Does the noise on a photo-diode signal follow a
Gaussian distribution? (10 marks)
As we discussed in Chapter 1, for very low intensities the distribution of counts from a photo-detector is
expected to follow the Poisson shot-noise distribution. However, for larger photon fluxes the noise on the
voltage generated in a photo-diode circuit is expected to follow a Gaussian distribution.
The above figure shows the signal output from a photo-diode as a function of time, and in part (b) a histogram
of the distribution of data. The number of observed data points lying within specified bands, , is given below.
Required:
(i) Use the equation,
to determine the number of data points expected in each interval, . You must give your
answer to 4 decimal places (2 marks)
# YOUR CODE HERE
2019/11/7 Assignment 4
https://notebooks.dmaitre.phyip3.dur.ac.uk/miscada-da/user/czjs88/notebooks/Homework 4/Assignment 4.ipynb 4/10
(ii) Show that for all bins, and state why there is no need to combine sequential bins. (2 marks)
(iii) Calculate from the formula (2 marks)
.
(iv) Calculate the number of degrees of freedom. (2 marks)
(v) Given the answers you found in (iii) and (iv), are the data consistent with the hypothesis of a Gaussian
distribution? (2 marks)
(i) Determine the number of data points expected in each interval, . You must
give your answer to 4 decimal places
Ei
In [ ]:
In [ ]:
(ii) Is for all bins? State why there is no need to combine sequential
bins.
Ei > 5
YOUR ANSWER HERE
(iii) Calculate χ .
2
In [ ]:
In [ ]:
(iv) What is the number of degrees of freedom?
def two_i():
intervals = [-1000000000,-2.5,-2,-1.5,-1,-0.5,0,0.5,1,1.5,2,2.5,10000000000]
expected_points = []
# YOUR CODE HERE
return(expected_points)
print(two_i())
'''TEST CELL- DO NOT DELETE'''
def two_iii():
# YOUR CODE HERE
return(chi2)
print(two_iii())
'''TEST CELL- DO NOT DELETE'''
2019/11/7 Assignment 4
https://notebooks.dmaitre.phyip3.dur.ac.uk/miscada-da/user/czjs88/notebooks/Homework 4/Assignment 4.ipynb 5/10
In [ ]:
In [ ]:
(v) Given the answers you found in (iii) and (iv), are the data consistent with the
hypothesis of a Gaussian distribution?
YOUR ANSWER HERE
Question 3: Benford's Law of Anomalous Numbers (10
marks)
It has been observed that the first pages of a table of common logarithms show more wear than do the last
pages, indicating that more used numbers begin with the digit 1 than with the digit 9. (F. Benford, Proceedings
of the American Philosophical Society, 78, March 1938). The Law can be stated that the probability of obtaining
a first digit is equal to .
Table 1 shows some of the data sets analysed by Benford, of which is saved as 'benford.csv'.
Let's consider the rivers row.
Required:
a log (1 + 1/a) 10
def two_iv():
# YOUR CODE HERE
return(degrees_of_freedom)
print(two_iv())
'''TEST CELL- DO NOT DELETE'''
2019/11/7 Assignment 4
https://notebooks.dmaitre.phyip3.dur.ac.uk/miscada-da/user/czjs88/notebooks/Homework 4/Assignment 4.ipynb 6/10
(i) Use Benford's Law to calculate the probability of obtaining a first digit of , for . (2
marks)
(ii) Using Benford's formula and the total count for that group, calculate the expected number
of occurrences of the first digit being , , for . (2 marks)
(iii) Ascertain whether some of the bins should be combined. (1 mark)
(iv) From the table, and the total count for that group, calculate the observed number of
occurrences of the first digit being , , for . (2 mark)
(v) Calculate and the number of degrees of freedom. (1 mark)
(vi) Test the hypothesis that the distribution for that group follows Benford's Law. (2 marks)
a 1 ≤ a ≤ 9
a Ea 1 ≤ a ≤ 9
a Oa 1 ≤ a ≤ 9
(i) Use Benford's Law to calculate the probability of obtaining a first digit of , for
(ii) Using Benford's formula and the total count for that group, calculate the
expected number of occurrences of the first digit being , , for
.
a Ea
(iii) Ascertain whether some of the bins should be combined.
YOUR ANSWER HERE
def three_i():
'''Your function must return an array of all the probabilities from a=1 to a=9'
probability = []
# YOUR CODE HERE
return(probability)
print(three_i())
'''TEST CELL- DO NOT DELETE'''
def three_ii():
expected_occurances = []
# YOUR CODE HERE
return(expected_occurances)
print(three_ii())
'''TEST CELL- DO NOT DELETE'''
2019/11/7 Assignment 4
https://notebooks.dmaitre.phyip3.dur.ac.uk/miscada-da/user/czjs88/notebooks/Homework 4/Assignment 4.ipynb 7/10
YOUR ANSWER HERE
(iv) From the table, and the total count for that group, calculate the observed
number of occurrences of the first digit being a, Oa, for 1 ≤ a ≤ 9.
(v) Calculate χ and the number of degrees of freedom.
(vi) Test the hypothesis that the distribution for that group follows Benford's Law
and make appropiate comments.
YOUR ANSWER HERE
Question 4: Analysing Bevington's data (10 marks)
In this question we will analyse the data from Table~8.1 in P.R. Bevington's book 'Data reduction and error
analysis'.
The data can be obtained from the file BevingtonData.csv. There are two columns: time, and number of counts
detected.
def three_iv():
'''Your function must return an array of occurances from a = 1 to a = 9'''
observed_occurances = []
# YOUR CODE HERE
return(observed_occurances)
print(three_iv())
'''TEST CELL- DO NOT DELETE'''
def three_v():
'''Your function must return the chi^2 value and the degrees of freedom'''
# YOUR CODE HERE
return(chi2,degrees_of_freedom)
print(three_v())
'''TEST CELL- DO NOT DELETE'''
'''TEST CELL- DO NOT DELETE'''
2019/11/7 Assignment 4
https://notebooks.dmaitre.phyip3.dur.ac.uk/miscada-da/user/czjs88/notebooks/Homework 4/Assignment 4.ipynb 8/10
The experiment involves irradiating a coin with thermal neutrons to create two short-lived silver isotopes that
subsequently decay by beta emission. Students count the emitted beta particles in 15s intervals for
approximately 4 minutes.
The model to describe the data has five parameters: a background, ; amplitudes of two excited states,
and , respectively; the mean lives of the excited states, and , respectively. Mathematically, we can
represent the decay by the fitting function:
Required:
(i) What are the count errors? (2 marks)
(ii) Perform a fit to find the parameters (with their errors) (4 marks)
(iii) What is the reduced value? (2 marks)
(iv) Plot an appropriate graph (i.e. counts versus time, with residuals) (2 marks)
(i) What are the count errors?
(ii) Perform a fit using the fitting function f defined in the cell below. What are the
best-fit parameters for a1, ..., a5 along with their errors?
data = pd.read_csv('BevingtonData.csv')
time = np.array(data.iloc[:,0])
count = np.array(data.iloc[:,1])
def count_errors():
'''Your function should return an array of the errors on each of the counts.'''
errors = []
# YOUR CODE HERE
return(errors)
print(count_errors())
'''TEST CELL- DO NOT DELETE'''
assert len(count_errors()) == 59, 'Please make sure that your function returns an ar
'''TEST CELL- DO NOT DELETE'''
2019/11/7 Assignment 4
https://notebooks.dmaitre.phyip3.dur.ac.uk/miscada-da/user/czjs88/notebooks/Homework 4/Assignment 4.ipynb 9/10
data = pd.read_csv('BevingtonData.csv')
time = np.array(data.iloc[:,0])
count = np.array(data.iloc[:,1])
def f(time,a_1,a_2,a_3,a_4,a_5):
return a_1 + a_2*np.exp(-time/a_4) + a_3*np.exp(-time/a_5)
def best_fit_line():
# YOUR CODE HERE
return(a1,a2,a3,a4,a5,error_a1,error_a2,error_a3,error_a4,error_a5)
print(best_fit_line())
'''TEST CELL- DO NOT DELETE'''
'''TEST CELL- DO NOT DELETE'''
'''TEST CELL- DO NOT DELETE'''
'''TEST CELL- DO NOT DELETE'''
'''TEST CELL- DO NOT DELETE'''
'''TEST CELL- DO NOT DELETE'''
'''TEST CELL- DO NOT DELETE'''
'''TEST CELL- DO NOT DELETE'''
2019/11/7 Assignment 4
https://notebooks.dmaitre.phyip3.dur.ac.uk/miscada-da/user/czjs88/notebooks/Homework 4/Assignment 4.ipynb 10/10
In [ ]:
In [ ]:
(iii) What is the reduced χ value?
2
In [ ]:
In [ ]:
(iv) Plot a suitable graph. This is, counts versus time, with residuals.
In [ ]:
'''TEST CELL- DO NOT DELETE'''
'''TEST CELL- DO NOT DELETE'''
def reduced_chi_squared():
reduced_chi_squared = 0
# YOUR CODE HERE
return(reduced_chi_squared)
'''TEST CELL- DO NOT DELETE'''
# YOUR CODE HERE