The parameter estimation and hypothesis test methods described above are statistically subordinate to the category of parameter statistics, and their common denominator is that the specific form of the distribution of the known population is assumed (such as the assumption that the population obeys the positive distribution), but there are some unknown parameters to be estimated or tested. We establish estimates or hypothesis test statistics based on extracted samples to obtain some inference of unknown parameters in the population. But in many cases we do not know the specific form of the overall distribution, arbitrarily assume that the overall compliance with a specific distribution is inevitably hasty, and if the actual overall distribution and the side of the hypothesis are very large differences, it is easy to produce the wrong conclusion.
Non-parametric inspection