1.  fweights, or frequency weights, are weights that indicate the number
        of duplicated observations.

    2.  pweights, or sampling weights, are weights that denote the inverse of
        the probability that the observation is included because of the
        sampling design.

    3.  aweights, or analytic weights, are weights that are inversely
        proportional to the variance of an observation; i.e., the variance of
        the jth observation is assumed to be sigma^2/w_j, where w_j are the
        weights.  Typically, the observations represent averages and the
        weights are the number of elements that gave rise to the average.
        For most Stata commands, the recorded scale of aweights is
        irrelevant; Stata internally rescales them to sum to N, the number of
        observations in your data, when it uses them.

    4.  iweights, or importance weights, are weights that indicate the
        "importance" of the observation in some vague sense.  iweights have
        no formal statistical definition; any command that supports iweights
        will define exactly how they are treated.  Usually, they are intended
        for use by programmers who want to produce a certain computation.

嗯     谢 谢 啊