np.random.choice () function
choice (a, size = None, replace = True, p = None): the given function from 1-dimensional array of random sampling.
- a: If a one-dimensional array, it means that from this one-dimensional array of random sampling; if int type, it represents a random sample from 0 to a-1 in this sequence.
- size: the number of sampling results, the default is 1 may be an integer representing the number to be sampled; may be a tuple, such as (m, n, k), the number of samples will have to m * n * k, size of ( m, n, k).
- replace: boolean type, sampling sample if you want to replace? I do not quite understand this place, test a little discovery replace designated to True, the sampling elements will be repeated; when replace designated as False, the sample will not be repeated.
- p: a one-dimensional array, the probability of developing a sample for each element, the same as if the default probability None, each element is a sampled.