Statistical sampling is a broad topic of research, but in practical machine learning, you may employ one of three types of sampling: simple random sample, systematic sampling, or stratified sampling. Simple Random Sampling: Samples are selected from the domain with a uniform probability.
The most common probability-based method is the stratified random sample, which divides a population into strata or groups before randomly selecting a subset from each group. The most common non-probability-based method is convenience sampling, which selects a sample from the population at convenience or random intervals.
There are four major methods: 1) basic random, 2) stratified random, 3) cluster, and 4) systematic. Non-probability sampling entails selecting the items of the sample using nonrandom means.
There are two kinds of sampling techniques: Probability sampling involves random selection, which allows you to draw strong statistical conclusions about the entire group. Non-probability sampling entails non-random selection based on convenience or other criteria, making it possible to collect data quickly.
Sampling is an important topic to understand in machine learning. It is important because without sampling, the model will not be able to learn anything. The model will never be able to generalize beyond the sample and it can’t make predictions on new data.