Source
Adapted from my own handwritten and hand-typed notes from Year 1 to Year 3 of my EEE degree
Estimation means inferring unknown population parameters (like and ) from known sample statistics (like and ).
Two types of estimate
Types:
- point estimate: a single number, for example km as the estimate of the mean distance
- interval estimate: a range believed to contain the parameter, for example km
A point estimate alone gives no idea of how close it is to the true value. An interval estimate fixes this by giving a range together with a level of confidence.
Estimators
The sample statistic used to estimate a parameter is called an estimator. The two main ones:
- estimates
- estimates
An estimator is unbiased if, averaged over many samples, it equals the parameter. Both and (with the divisor) are unbiased.
Confidence interval idea
A confidence interval is an interval estimate with a confidence level (for example 90% or 95%).
"The confidence is 90% that contains " means: if we took many samples and built an interval from each in the same way, about 90% of these intervals would contain .
General notation used in the course:
- confidence level
- each tail gets area
Higher confidence gives a wider interval. A larger sample gives a narrower interval. The working method is in Confidence Interval for the Mean.