Advantages: simple operation;
Disadvantages: there is no guarantee that the sample can perfectly represent the whole population;
Applicable: Scenes where individuals are evenly distributed.
2) equidistant sampling: firstly, number each individual in the group in sequence, then calculate the sampling interval, and then select individuals according to a fixed number.
Advantages: simple operation;
Disadvantages: when the distribution law is obvious, it is easy to produce deviation;
Applicability: Scenes with uniform distribution of individuals show obvious uniform distribution law.
3) Stratified sampling: firstly, all individual samples are divided into several categories according to certain characteristics, and then individuals are selected from each category by random sampling or equidistant sampling.
Advantages: reducing sampling error and conducting separate research on different types of data samples;
Disadvantages: no shortcomings;
Applicability: Data with attributes and labels of classification logic.
4) Cluster sampling: first, divide all samples into several small population sets, and then randomly select several small population sets to represent the population.
Advantages: simple operation;
Disadvantages: the distribution is limited by the division of small groups, and the sampling error is large;
Applicability: The characteristics of small group sets are relatively small, and there are higher requirements for dividing small group sets.