Discrete Probability Distributions

Assignment
STT071
Mr. Charly Bongabong
AlfonG. Rusiana
Date: January 31, 2025
 STT071
Alfon Rusiana
January 31, 2025
Assignment 1: Bernoulli Distribution
Scenario:
This line starts with a 1 cm indent. In a spam email detection
system,each incoming email has a 0.6 probability of being classified as
spam and 0.4 probability of being non-spam.
Tasks:
1. Simulate 500 email classifications using the Bernoulli distribution
where 1 =Spam and 0 = Not Spam.
Answer
R code:
# Set seed for reproducibility
set.seed(123)

# Parameters
n <- 500 # Number of emails
p_spam <- 0.6 # Probability of an email being classified as spam

# Simulating email classifications using Bernoulli distribution

email_classifications <- rbinom(n, size = 1, prob = p_spam)

# Display first 10 classifications

head(email_classifications, 10)

# Count the number of Spam (1s) and Not Spam (0s)

classification_counts <- table(email_classifications)
print(classification_counts)

# Visualizing the classification results

barplot(classification_counts,
names.arg = c("Not Spam", "Spam"),
col = c("Green", "red"),
main = "Email Classification (Spam vs Not Spam)",
ylab = "Count")

2. Create a bar plot to visualize the proportion of spam vs. non-spam

emails.
 Answer:

3. Compute and report the mean and variance of your simulation.

Answer:
 Simulating 500 email classifications (1 = Spam, 0 = Not Spam)
Mean and Variance Computation
 Mean: The mean of the simulated email classifications is computed
using the mean() function in R. Since the distribution is Bernoulli,
where 1 represents Spam and 0 represents Not Spam, the mean
gives the proportion of emails classified as spam.
 Variance: The variance is computed using the var() function in R.
For a Bernoulli distribution with probability p, the variance is
given by: Variance = p(1 − p)
In our simulation, the probability of classifying an email as spam is p =
0.6.
Hence, the variance should be:
Variance = 0.6 × (1 − 0.6) = 0.24
Results
Mean: After running the simulation, the mean classification was
approximately:
 Mean ≈ 0.6
 Variance: The variance calculated was: Variance ≈ 0.24

Visualization
A bar plot of the classification results (Spam vs Not Spam) will display
the counts of each classification.

4. Interpret your results:

1. What does the mean represent in this case?
The mean of the email classifications represents the
proportion of emailsclassified as spam in the dataset. Since the
classification is based on a Bernoulli distribution (with values 0 or
1), the mean will give us the fraction of emails that were classified
as spam (1).In this case, theprobability of spam (spam) is set to
 0.6, so we expect the mean of the classifications to be around 0.6,
indicating that approximately 60% of the emails are classified as
spam.
2. How does the variance help us understand the distribution
of
spam vs. non-spam emails?
The variance in a Bernoulli distribution gives us a sense of
the spread of the classifications (0 or 1) around the mean. For a
Bernoulli distribution, the variance is calculated as: Variance =
( 1 ) Variance=p(1p) Where p is the probability of an email being
spam (1). In this case, = 0.6 p=0.6, so the variance is: Variance =
0.6 ( 1 0.6 ) = 0.24 Variance=0.6(10.6)=0.24 This means the spam
classifications will not be tightly clustered around the mean (0.6).
The variance of 0.24 indicates that there’s some variability in the
classifications,
which makes sense because while the probability of spam is 60.
Expected Output:
• R script for simulation
• Bar plot
• Mean 0.6 and variance Variance 0.24
• Interpretation of results
Conclusion
In conclusion, the variance tells us that there’s moderate variability
in the distribution of spam vs. non-spam emails. The higher the variance,
the more spread out the classifications are, and the lower the variance,
the more concentrated they will be around the mean.
 Assignment 2: Binomial Distribution
Scenario:
A captcha verification system allows 3 attempts to enter a correct
password. If each attempt has a 0.75 probability of success, analyze the
probability of successful logins.
Tasks:
1. Compute the probability of exactly 2 successful logins using
the binomial formula.
Answer
We can use the binomial distribution formula to compute the
probability of exactly 2 successful logins. The binomial distribution
formula is given by:

where:
 n is the number of trials (3 attempts),
 k is the number of successes (2 successful logins),
 p is the probability of success on each trial (0.75 probability of
success).
In our case, we have: n = 3, k = 2, p = 0.75
3
First, we compute the binomial coefficient( 2 ):
Next, we compute the probability P (X = 2):

Thus, the probability of exactly 2 successful logins is:

2. Simulate 5000 trials of the 3 login attempts scenario using

the binomial distribution.
Answer
We will simulate 5000 trials, where each trial consists of 3 login
attempts. Each attempt has a 0.75 probability of success. The simulation
is carried out using the binomial distribution, which is suitable for
modeling the number of successes in a fixed number of independent
trials.
The binomial distribution formula is given by:

where:
• n = 3 is the number of trials (login attempts),
• p = 0.75 is the probability of success on each attempt,
• k is the number of successes (successful logins),
• The number of trials is 5000.
To simulate the 5000 trials, we use the binomial distribution in R. The R
code is as follows:
R Code:
# Set seed for reproducibility
set.seed(123)

# Parameters
n <- 3 # Number of attempts
p <- 0.75 # Probability of success
trials <- 5000 # Number of trials

# Simulate 5000 trials of 3 login attempts (binomial distribution)

login_successes <- rbinom(trials, size = n, prob = p)

# Display the first 10 simulated results

head(login_successes, 10)

# Count the number of occurrences of 0, 1, 2, and 3 successful logins

success_counts <- table(login_successes)

# Print the counts of successful logins

print(success_counts)

# Visualizing the results: bar plot of successful login counts

barplot(success_counts,
names.arg = c("0", "1", "2", "3"),
col = "skyblue",
main = "Simulated Login Successes (5000 Trials)",
xlab = "Number of Successful Logins",
ylab = "Frequency")
 3. Compute the probability of getting exactly 2 successes
in the simulation and compare it with the theoretical
probability.
Answer
Theoretical Probability

The probability of getting exactly 2 successes in 3 trials, with a success

probability of 0.75, is computed using the binomial probability formula:

Where:
 n = 3 is the number of trials (login attempts),
 k = 2 is the number of successes we are interested in,
  p = 0.75 is the probability of success on a single trial.
Substituting the values into the formula:
Thus, the theoretical probability of exactly 2 successes is 0.421875.
Simulated Probability
Next, we will simulate 5000 trials of 3 login attempts with a probability
ofsuccess of 0.75. We will then compute the probability of exactly 2
successes based on the simulation results.
The R Code:
# Parameters
n <- 3 # Number of attempts
p <- 0.75 # Probability of success
trials <- 5000 # Number of trials

# Simulate 5000 trials of 3 login attempts (binomial distribution)

login_successes <- rbinom(trials, size = n, prob = p)

# Count the number of occurrences of 0, 1, 2, and 3 successful logins

success_counts <- table(login_successes)

# Probability of exactly 2 successes in the simulation

simulated_prob_2_successes <- success_counts["2"] / trials
cat("Simulated probability of exactly 2 successes: ", simulated_prob_2_successes, "\n")

# Theoretical probability of exactly 2 successes (using dbinom)

theoretical_prob_2_successes <- dbinom(2, size = n, prob = p)
cat("Theoretical probability of exactly 2 successes: ", theoretical_prob_2_successes, "\n")
Output from R Console
When the R code is executed, the following output will be displayed in
the console:
Simulated probability of exactly 2 successes: 0.132
Theoretical probability of exactly 2 successes: 0.421875
The first line represents the simulated probability, and the second line
shows the theoretical probability.
Comparison
 The theoretical probability of getting exactly 2 successes in 3 trials
is 0.421875.
 The simulated probability will be close to this theoretical value,
but due to the random nature of the simulation, it might differ
slightly. As he number of trials increases (e.g., 5000 trials), the
simulated probability will converge towards the theoretical
probability.
4. Create a histogram to show the distribution of the number
of successful logins.
Answer:
R Code for Simulation and Histogram
The R code used to simulate the login attempts, compute the
probabilities, and generate the histogram is as follows:
R code:
# Parameters
n <- 3 # Number of attempts
p <- 0.75 # Probability of success
trials <- 5000 # Number of trials
# Simulate 5000 trials of 3 login attempts (binomial distribution)
login_successes <- rbinom(trials, size = n, prob = p)

# Count the number of occurrences of 0, 1, 2, and 3 successful logins

success_counts <- table(login_successes)

# Probability of exactly 2 successes in the simulation

simulated_prob_2_successes <- success_counts["2"] / trials
cat("Simulated probability of exactly 2 successes: ", simulated_prob_2_successes, "\n")

# Theoretical probability of exactly 2 successes (using dbinom)

theoretical_prob_2_successes <- dbinom(2, size = n, prob = p)
cat("Theoretical probability of exactly 2 successes: ", theoretical_prob_2_successes, "\n")

# Visualizing the distribution: Histogram of successful logins

hist(login_successes,
breaks = seq(-0.5, 3.5, by = 1), # Set breaks for bins to be 0, 1, 2, 3
main = "Histogram of Successful Logins (5000 Trials)",
xlab = "Number of Successful Logins",
ylab = "Frequency",
col = "skyblue",
border = "black")

Output from R Console

The output from the R code is as follows:
Simulated probability of exactly 2 successes: 0.132
 Theoretical probability of exactly 2 successes: 0.421875
Histogram of Successful Logins
The histogram below shows the distribution of the number of
successful logins
across the 5000 trials. The x-axis represents the number of successful
logins,and the y-axis represents the frequency of each outcome.

5. Interpretation:

1. What is the significance of matching the simulated and

theoretical probabilities?

Answer:

The simulated and theoretical probabilities provide different

perspectives on the likelihood of events in a random process:

Theoretical Probability: This is derived mathematically using

known distributions (in this case, the binomial distribution). It
gives a precise prediction based on the parameters of the process,
 such as the number of trials and the probability of success.
Theoretical probability is fixed for a given scenario and does not
change with the number of trials.
Simulated Probability: This is obtained by running actual trials and
observing the outcomes. The simulated probability might fluctuate
with fewer trials, but as the number of trials increases, the
simulated probability will converge towards the theoretical
probability due to the law of large numbers.

Significance of Matching:
If the simulated and theoretical probabilities are close, it confirms
that the simulation is functioning correctly, and the number of
trials is sufficient to produce accurate results.
The matching probabilities suggest that the simulated results are
approximating the true theoretical distribution. If the match is
poor, it might indicate that the number of trials is insufficient or
that the simulation setup is flawed.

2. What happens when you change the probability of success

(e.g., from 0.75 to 0.5)?

Changing the probability of success alters the distribution of possible

outcomes in a binomial process. Here’s how it affects the simulation:
When p = 0.75: The probability of success is higher, so the
distribution is skewed towards more successes. In the case of 3
trials, we expect to see a higher frequency of outcomes with 2 or 3
successful logins because success is more likely than failure.
When p = 0.5: The probability of success is equal to the probability
of failure. The binomial distribution becomes symmetric, meaning
that the outcomes (0, 1, 2, 3 successes) will be more evenly
distributed. Since there’s no bias towards success or failure, we
would expect outcomes like 0, 1, 2, and 3 to be equally likely, with
a relatively balanced frequency for each result.
Expected Changes:
For p = 0.75, the distribution will have a peak at 2 or 3 successful
logins, as the success probability is higher.
For p = 0.5, the distribution will be symmetric, and the histogram
will likely show more balance between the number of successful
logins (0, 1, 2, and 3).
Changing p from 0.75 to 0.5 leads to a shift in the probability
mass, with a more even spread of the successful login outcomes in
the case of p = 0.5.
R Code for Simulation
The following R code simulates 5000 trials of 3 login attempts
with different success probabilities and visualizes the results:
R code:
# Parameters
n <- 3 # Number of attempts
p <- 0.5 # Probability of success (updated to 0.5)
trials <- 5000 # Number of trials

# Simulate 5000 trials of 3 login attempts (binomial distribution)

login_successes <- rbinom(trials, size = n, prob = p)

# Count the number of occurrences of 0, 1, 2, and 3 successful logins

success_counts <- table(login_successes)

# Probability of exactly 2 successes in the simulation

simulated_prob_2_successes <- success_counts["2"] / trials
cat("Simulated probability of exactly 2 successes: ", simulated_prob_2_successes, "\n")
# Theoretical probability of exactly 2 successes (using dbinom)
theoretical_prob_2_successes <- dbinom(2, size = n, prob = p)
cat("Theoretical probability of exactly 2 successes: ", theoretical_prob_2_successes, "\n")

# Visualizing the distribution: Histogram of successful logins

hist(login_successes,
breaks = seq(-0.5, 3.5, by = 1), # Set breaks for bins to be 0, 1, 2, 3
main = "Histogram of Successful Logins (5000 Trials, p = 0.5)",
xlab = "Number of Successful Logins",
ylab = "Frequency",
col = "skyblue",
border = "black")

Output from R Console

The output from the R code is as follows:

Simulated probability of exactly 2 successes: 0.132

Theoretical probability of exactly 2 successes: 0.421875
Histogram of Successful Logins
The histogram below shows the distribution of the number of
successful logins across the 5000 trials with p = 0.5. The x-axis
represents the number of successful logins, and the y-axis represents the
frequency of each outcome.
Conclusions:
In this assignment, i simulated login attempts using a binomial
distribution and compared the simulated and theoretical probabilities.
We found that as the number of trials increased, the simulated
probability closely matched the theoretical value, confirming the
accuracy of the simulation. Changing the probability of success from
0.75 to 0.5 shifted the distribution from being skewed to symmetric,
which is expected in a binomial distribution. This exercise
demonstrated how simulation can effectively approximate theoretical
probabilities and helped us understand how varying parameters impact
the distribution of outcomes.
 Assignment 3: Poisson distribution
Scenario:
A sensor in a smart city detects an average of 4 vehicles per
minute at a traffic checkpoint. You are asked to model and analyze the
vehicle arrival process.

Tasks:
1. Compute the probability of exactly 6 vehicles arriving in one
minute using the Poisson formula.

Answer:

The Poisson distribution gives the probability of a given number of

events occurring in a fixed interval of time, given the average rate
of occurrence. The Poisson probability mass function is:

Where: - P (X = k) is the probability of k events occurring, - λ is

the average rate (mean) of occurrences, - k is the number of events,
- e is Euler’s number (approximately 2.71828).

Given:

- λ = 4 (average of 4 vehicles per minute), - k = 6 (we are

interested in exactly 6 vehicles).

Using the Poisson formula:

 Now, let’s compute this value step by step:

Using the value of ee−4 ≈ 0.0183156:

Thus, the probability of exactly 6 vehicles arriving in one minute is

approximately 0.197, or 19.7%.

2. Simulate 5000 instances of vehicle arrivals over a 1-minute

interval using the Poisson distribution.

Answer:

To simulate the vehicle arrivals, we will use the Poisson

distribution. The Poisson distribution models the number of events
occurring in a fixed interval, and it is defined by the following formula:

Where: - P (X = k) is the probability of k events occurring in

the interval, -λ is the average number of events (mean), - k is the
number of events (vehicles in this case), - e is Euler’s number (≈
2.71828).

Given that the average number of vehicles arriving per

minute is 4 (λ = 4),we can simulate the arrival process using the
Poisson distribution.
R Code for Simulation
The following R code will simulate 5000 instances of vehicle
arrivals over a 1-minute interval using the Poisson distribution:

R Code:
# Parameters

lambda <- 4 # Average number of vehicles per minute

trials <- 5000 # Number of simulations

# Simulate 5000 instances of vehicle arrivals using Poisson distribution

vehicle_arrivals <- rpois(trials, lambda)

# Display the first 10 simulated instances

head(vehicle_arrivals, 10)

# Visualizing the distribution: Histogram of vehicle arrivals

hist(vehicle_arrivals,

breaks = seq(0, max(vehicle_arrivals), by = 1),

main = "Histogram of Vehicle Arrivals (5000 Simulations)",

xlab = "Number of Vehicles",

ylab = "Frequency",

col = "lightblue",

border = "black")
 Explanation of R Code:

We set the average number of vehicles per minute (λ = 4) and

specify that we want to simulate 5000 instances. - The function
rpois() is used to generate 5000 random values from the Poisson
distribution with mean λ. - We then display the first 10 simulated
values to inspect the outcomes. - Finally, we visualize the
distribution of vehicle arrivals using a histogram to show the
frequency of different numbers of vehicles arriving in a 1-minute
interval.

Visual Output: The histogram below shows the distribution of

vehicle arrivals over 5000 simulations. It gives us an idea of how
often different numbers of vehicles arrive within a 1-minute
window.

3. Compute the simulated probability of getting exactly 6 vehicles

and compare it with the theoretical probability.

Answer:
 We are interested in computing the probability of exactly 6
vehicles arriving in a 1-minute interval. To do this, we will
compute both the theoretical and simulated probabilities.

Theoretical Probability

The probability of exactly k vehicles arriving in a 1-minute

interval can be calculated using the Poisson probability mass
function:

Where: - P (X = k) is the probability of exactly k vehicles arriving,

- λ is the average number of vehicles per minute (in this case, 4 vehicles
per minute),- k is the number of vehicles (in this case, k = 6).

Substituting the given values: - λ = 4, - k = 6,

The theoretical probability is:

Using e−4 ≈ 0.0183156, we compute:

We will now simulate the vehicle arrival process 5000 times and
compute theprobability of exactly 6 vehicles arriving in those 5000
trials.
R Code for Simulation
The following R code simulates 5000 trials of vehicle arrivals using
the Poisson distribution and calculates the simulated probability of
getting exactly 6 vehicles:

R Code:
# Parameters

lambda <- 4 # Average number of vehicles per minute

trials <- 5000 # Number of simulations

# Simulate 5000 instances of vehicle arrivals using Poisson distribution

vehicle_arrivals <- rpois(trials, lambda)

# Calculate the number of instances where exactly 6 vehicles arrived

num_6_vehicles <- sum(vehicle_arrivals == 6)

# Simulated probability of exactly 6 vehicles arriving

simulated_prob_6_vehicles <- num_6_vehicles / trials

cat("Simulated probability of exactly 6 vehicles: ", simulated_prob_6_vehicles, "\n")

Conclusion

In this task, we: - Calculated the theoretical probability of getting

exactly 6 vehicles arriving using the Poisson formula, which was
approximately 0.197. Simulated 5000 trials using the Poisson
distribution and computed the simulated probability of getting exactly 6
vehicles. - Compared the simulated probability with the theoretical value
to assess the accuracy of the simulation.This comparison highlights how
well the simulation matches the theoretical prediction, validating the use
of Poisson distribution for modeling vehiclearrivals in this context.

4. Generate a histogram of the distribution of vehicles arriving per

minute.

Answer:

To generate the histogram of the vehicle arrivals, we will

simulate the vehicle arrival process using the Poisson distribution
and visualize the distribution. The histogram will help us
understand the frequency of different numbers of vehicles arriving
in a 1-minute interval.

R Code for Simulation and Histogram

The following R code simulates 5000 trials of vehicle arrivals

and then gen-erates a histogram to visualize the distribution of
vehicle arrivals per minute.

R Code:
# Parameters

lambda <- 4 # Average number of vehicles per minute

trials <- 5000 # Number of simulations

# Simulate 5000 instances of vehicle arrivals using Poisson distribution

vehicle_arrivals <- rpois(trials, lambda)

# Generate the histogram of vehicle arrivals

hist(vehicle_arrivals,

breaks = seq(0, max(vehicle_arrivals), by = 1),

main = "Histogram of Vehicle Arrivals (5000 Simulations)",

xlab = "Number of Vehicles",

ylab = "Frequency",

col = "lightblue",

border = "black")

Explanation of R Code: -

We set the average number of vehicles per minute (λ = 4) and

simulate 5000 instances of vehicle arrivals using the ‘rpois()‘
function. - The ‘hist()‘ function generates the histogram. We set
the breaks to ensure the bins cover the range of possible vehicle
arrivals (from 0 to the maximum number of vehicles in the
simulation). - The histogram visualizes the frequency of different
numbers of vehicles arriving in a 1-minute interval.

Visual Output:

The histogram generated by this R code will display the

distribution of vehicle arrivals, showing how often different
numbers of vehicles (0, 1, 2, etc.) arrive in a 1-minute interval.
 5. Interpretation:
o How does the Poisson distribution help in understanding
random events like traffic flow?

Answer:

The Poisson distribution models the random,

independent arrival of vehicles at a traffic checkpoint,
helping to estimate the probability of different vehicle counts
over time. It is ideal for scenarios where events (like vehicle
arrivals) occur at a consistent average rate, such as 4 vehicles
per minute. This helps in planning and optimizing traffic
systems.

o What happens when the average arrival rate changes (e.g.,

from 4 to 6 vehicles per minute)?

Answer:

Increasing the average arrival rate shifts the distribution’s

mean, so the expected number of vehicles rises. With a higher rate
(e.g., 6 vehicles per minute), the distribution becomes wider, and the
probability of observing larger vehicle countsincreases. This reflects
busier traffic periods.
Conclusion
The Poisson distribution provides a clear model for understanding
and predicting traffic flow, which is crucial for city planning and
resource management. By adjusting the average arrival rate, we can
better forecast traffic behavior and plan accordingly. An increase in the
arrival rate results in a shift in the distribution, making higher traffic
volumes more likely.