Chapter 2: Overview of each question
2.1 Bayes’ Theorem for Conditional Probability (Problem 1, 6, 8, 9):
Formula:
P(A ∣ B) =
P (B ∣ A) . P ( A )
P(B)
Bayes' Theorem is ideal for calculating the probability of a cause given an observed
effect, making it perfect for scenarios with multiple hypotheses and conditional
dependencies. In Problem 1, it determines the probability a defective item comes from
Supplier X by combining defect rates and selection probabilities. Similarly, it’s used in
Problem 6 for spam email classification, Problem 8 for handwriting recognition, and
Problem 9 for error detection in communication systems. The formula’s strength lies in
updating probabilities with new evidence, crucial for diagnostic and classification tasks.
It handles scenarios where prior probabilities (e.g., supplier selection, spam likelihood)
and conditional probabilities (e.g., defect rates, word occurrences) need to be integrated
to assess risk or make decisions, providing a robust framework for quality control and
machine learning.
External Applications:
Medical Diagnostics: Bayes' Theorem is used to calculate the probability of a disease
given a positive test result, accounting for test sensitivity and disease prevalence, as in
cancer screening or COVID-19 testing.
Fraud Detection: Financial institutions apply Bayes’ Theorem to assess the likelihood of
fraudulent transactions based on patterns like unusual spending or location data.
Autonomous Vehicles: It helps in sensor fusion, combining data from cameras and radar
to estimate the probability of obstacles, enhancing navigation safety.
2.2 Binomial Distribution (Problem 2, 4):
Formula:
n
P(X = x) = ( x ) . p x . (1− p)n− x
The binomial distribution models the number of successes (e.g., defective items) in a
fixed number of independent trials with a constant success probability. In Problems 2 and
4, it calculates the probability of exactly 2 defective microchips and at least 1 defective
item in a sample of 10, respectively. Its discrete, binary nature fits quality control
scenarios where outcomes are defective or non-defective. It provides a precise way to
evaluate the likelihood of specific defect counts, enabling engineers to assess production
reliability and determine if processes require further inspection, critical for maintaining
quality standards.
External Applications:
Pharmaceutical Testing: Used to evaluate the probability of side effects in clinical trials,
such as the number of patients experiencing adverse reactions in a sample.
Network Reliability: Applied to assess the likelihood of packet loss in communication
networks, where each packet has a fixed probability of failure.
Marketing Campaigns: Helps predict the success rate of customer responses (e.g., clicks
on an ad) in a fixed sample, aiding campaign optimization.
2.3 Normal Distribution (Problem 3):
Formula:
X−μ
Z=
σ
The normal distribution models continuous variables like daily demand in Problem 3,
where demand follows a bell-shaped curve. It calculates the safety stock needed to limit
stockout risk to 5%, using the Z-score to find the threshold for a 95% probability, critical
for inventory management. Its symmetry and well-tabulated Z-values allow precise
determination of stock levels to balance customer service and inventory costs, making it
ideal for demand forecasting in supply chain operations.
External Applications:
Financial Risk Management: Used to model stock price fluctuations or portfolio returns,
helping set risk thresholds for investments.
Weather Forecasting: Applied to predict temperature or rainfall distributions, aiding in
resource planning for agriculture or disaster preparedness.
Manufacturing Tolerance Analysis: Helps determine acceptable ranges for product
dimensions, ensuring components meet specifications.
2.4 Complement Rule (Problem 4, 5, 7)
Formula:
P(at least 1 event) = 1 – P(nothing occur)
This method simplifies calculating the probability of at least one occurrence, such as
defective items (Problem 4), supplier delays (Problem 5), or hash table collisions
(Problem 7). In Problem 5, it computes the probability of at least one supplier delay by
calculating the probability all suppliers are on time. In Problem 7, it determines the
likelihood of at least one hash table collision. This approach is efficient for independent
events. By focusing on the complement (no events), it reduces computational complexity,
making it practical for risk analysis and system performance evaluation in diverse
engineering contexts.
External Applications:
Cybersecurity: Used to calculate the probability of at least one security breach in a
network with multiple entry points, aiding in defense strategy planning.
Project Management: Helps estimate the likelihood of at least one task delay in a project
with multiple independent tasks, informing timeline adjustments.
Reliability Engineering: Applied to assess the probability of at least one component
failure in a system with multiple parts, guiding maintenance schedules.
2.5 Hash table collision probability (Problem 7)
Formula:
n n−1 n−2 n−k + 1
P(no collision) = . . …
n n n n
Or:
n!
P(no collision) =
( n−k ) ! .n k
This combinatorial probability formula is tailored for analyzing hash table performance
under uniform hashing. Problem 7 calculates the probability of at least one collision by
determining the likelihood that all 5 keys hash to distinct buckets and subtracting from 1.
It accounts for the decreasing number of available buckets as keys are assigned.
Collisions impact hash table efficiency, and this formula provides a precise way to
evaluate performance, critical for designing data structures in computer engineering
applications like databases or caching systems.
External Applications:
Database Optimization: Used to assess collision probabilities in database indexing,
ensuring efficient query performance in systems like SQL databases.
Caching Systems: Helps evaluate the likelihood of cache conflicts in web servers,
optimizing content delivery in platforms like CDNs.
Cryptographic Hashing: Applied to analyze collision risks in hash functions used for data
integrity, such as in blockchain or digital signatures.