For example, we need to estimate parameter a based on a series of samples.

Then, we can define such a quantity: it is represented by a, but its distribution does not depend on a, and we call this quantity fulcrum quantity.

For example, if A is the mean of a normal distribution with known variance and the sample mean is 0, then following the known normal distribution, we can call B the central quantity.

It is easy to see that the hub quantity has two properties: 1. The distribution is known, and 2. It contains information about unknown parameters.

We write the estimated principal component of A as f(a, x), where x represents the sample. Because the distribution of fulcrum quantity is known, it is possible for us to find such an interval [bl, bh] that the probability is greater than 95%. Further, if the inequality equivalent to inequality can be found, it can be concluded that the probability that A falls within the interval is not less than 95%, that is, the interval is a confidence interval with 95% confidence.