Scribd
Upload
23 views2 pages

Bayesian 1 - Exercise Three

The document outlines various exercises related to Bayesian statistics, including the computation of posterior distributions and Bayes estimators under different loss functions. It covers scenarios involving defective items, coin tosses, electronic component lifetimes, fertilizer effects on crop yield, matatu reliability, and traffic accident modeling. Each exercise requires the application of Bayesian methods with specified prior distributions and observed data.

Uploaded by

seniobezo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
23 views2 pages

Bayesian 1 - Exercise Three

The document outlines various exercises related to Bayesian statistics, including the computation of posterior distributions and Bayes estimators under different loss functions. It covers scenarios involving defective items, coin tosses, electronic component lifetimes, fertilizer effects on crop yield, matatu reliability, and traffic accident modeling. Each exercise requires the application of Bayesian methods with specified prior distributions and observed data.

Uploaded by

seniobezo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Exercise Three

1. Assume that you want to investigate the proportion 𝜃 of defective items


manufactured at a production line. Your colleague takes a random sample of 30
items. There were three defective items in the sample. Assume a uniform prior for 𝜃.
Compute the posterior distribution for 𝜃. Find the bayes estimator under the
quadratic loss function.
2. Suppose you observe 14 heads in 17 tosses of an unbiased coin. Assume the prior is
Beta prior with parameters 𝛼 = 𝛽 = 1. Find the posterior distribution. Find the bayes
estimator under the quadratic loss function.
3. Consider the lifetimes of Electronic Components. Suppose that we observe the
lifetimes of three components, 𝑋1 = 3, 𝑋2 = 1.5, 𝑎𝑛𝑑 𝑋3 = 2.1. These were
modeled as i.i.d. exponential random variables given 𝜃. Our prior distribution for
𝜃 is the gamma distribution with parameters 1 and 2.
a) What is the posterior distribution of θ given these observed lifetimes?
b) Determine the bayes estimator under the squared error loss function.
4. A large shipment of parts is received, out of which 9 are tested for defects. From past
shipments, it is known that 𝜃 , the probability that a part chosen at random is
defective, has a 𝐵𝑒𝑡𝑎 (1, 9) prior distribution. Find the;
a) Posterior distribution for 𝜃.
b) If 𝑥 = 0 is observed find
i) Bayes estimator under Quadratic loss function
ii) Bayes estimator under Absolute error loss function
5. A farmer wants to determine whether a new type of fertilizer improves crop yield.
The farm is divided into 10 plots, and each plot is randomly assigned either the new
fertilizer or the standard fertilizer. The outcomes (𝑌) represent whether the yield
from a plot is "improved" (𝑌 = 1) or "not improved" (𝑌 = 0).
The farmer observes the following data: New Fertilizer: 6 plots showed
improvement (𝑌 = 1) and 4 plots did not (𝑌 = 0). Assume that the probability of
improvement with the new fertilizer is 𝜃 ∈ [0,1]. The farmer uses a Beta prior
distribution, 𝑃(𝜃) = 𝐵𝑒𝑡𝑎(𝛼, 𝛽), with parameters α=2 and β=2
a) Derive the posterior distribution for 𝜃.
b) Compute the posterior mean of 𝜃.
6. Rose joined Kirinyaga University and could not get accommodation in campus. She
rented a house far away from campus. Five matatus were available to take her to
campus. At first she took any matatu that came along and was heading in the right
direction. Of the five matatus, two of them took her to the right place but the other
three far from work leaving her with 30 minutes to walk. Suppose you would want
to infer the proportion 𝜃, of the matatus that are “good”, that would take her right to
campus.
Assume that the set of possible values for 𝜃 is:
{0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1}
Assume also that before we got the data (2 successes out of 5 trials), we were very
uncertain about the value of 𝜃, and this can be modelled by using a uniform prior
1
distribution such that the probability of each is 11 .
a) Compute the Posterior probabilities.
b) Find the Bayes estimator for 𝜃 under the quadratic loss function.
7. A transportation engineer studying traffic accidents in the Kenyan Central Region
models the number of accidents per day as independent Poisson random variables
with unknown parameter 𝜆 accidents/day. 𝑋𝑖 ′𝑠 represents the number of accidents
in the ith day of observation. The engineer has chosen a conjugate prior Gamma of
unknown parameter 𝜆 with parameters shape 𝛼 = 2 and scale 𝛽 = 0.48. The data he
observed was 6 one-hour observation periods with:
1, 0, 1, 2, 1, 0 errors.
For the unknown parameter 𝜆, find the:
a) Posterior distribution
b) Bayes estimate under the quadratic loss function.
c) Posterior standard deviation

You might also like