📘 Confidence Intervals — Concept & Interpretation

Instead of giving a single estimate, confidence intervals provide a range that accounts for sampling uncertainty.

🎯 Why Confidence Intervals Are Needed

Point estimates vary from sample to sample due to natural randomness.

Therefore, a single estimate cannot fully represent the true population value.

📦 Components of a Confidence Interval

A confidence interval consists of three important parts:

1️⃣ Point Estimate

The best single estimate obtained from sample data.

Examples: sample mean (x̄), sample proportion (p̂)

2️⃣ Margin of Error (ME)

The amount added and subtracted from the point estimate to create a range.

3️⃣ Confidence Level

The probability that the interval estimation method captures the true population parameter.

📐 General Form of Confidence Interval

Confidence Interval = Point Estimate ± Margin of Error

\[ \text{CI} = \text{Estimate} \pm \text{ME} \]

This creates a lower limit and an upper limit.

🔢 Example: Estimating Average Exam Score

Suppose a sample of students gives:

• Sample mean = 72 marks
• Margin of error = 3 marks

Confidence interval:

72 ± 3

We estimate that the population mean lies between 69 and 75.

🎚️ Understanding Confidence Level

Common confidence levels are:

• 90% Confidence Level
• 95% Confidence Level
• 99% Confidence Level

📊 Interpretation Example

A 95% confidence interval for mean height is (168 cm, 172 cm).

Correct Interpretation:

Incorrect Interpretation:

• The population mean has a 95% probability of being in the interval ❌
• 95% of individual heights lie in the interval ❌

⚖️ Factors Affecting Confidence Interval Width

1️⃣ Sample Size (n)

• Larger samples → Narrower intervals
• Smaller samples → Wider intervals

2️⃣ Variability (σ)

• Higher variability → Wider intervals
• Lower variability → Narrower intervals

3️⃣ Confidence Level

• Higher confidence → Wider interval
• Lower confidence → Narrower interval

📈 Visual Intuition

Imagine repeatedly sampling and computing intervals:

• Most intervals contain the true value
• A few miss due to randomness

Confidence level measures the success rate of this method.

🧠 Why Confidence Intervals Are Better Than Point Estimates

🤖 Importance in Machine Learning

• Evaluating model accuracy ranges
• Estimating error margins
• Comparing model performances
• Reliability of predictions

🧠 Key Insights

• Confidence intervals provide ranges, not exact values
• They combine estimation and uncertainty
• Higher confidence produces wider intervals
• Larger samples produce more precise estimates
• They form the basis for statistical decision-making