statistickway.py [9:21]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
def get_sensitivity(numqueries):
    def get_sensitity_k(D):
        n, d = D.shape
        return sqrt(numqueries) / n

    return get_sensitity_k


# k way correlations --- i.e. for all k-tuple q in queries we compute the value \sum_t (\prod_j D[t,j], j in q)
def preserve_statistic(queries):

    letters_idx = string.ascii_lowercase
    row_idx, col_idxs = letters_idx[0], letters_idx[1:]
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



statistickway_threshold.py [9:22]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
def get_sensitivity(numqueries):
    def get_sensitity_k(D):
        n, d = D.shape
        return sqrt(numqueries) / n

    return get_sensitity_k


# k way correlations --- i.e. for all k-tuple q in queries we compute the value \sum_t (\prod_j D[t,j], j in q)
def preserve_statistic(queries):
    # hardcoding threshold to 1 here

    letters_idx = string.ascii_lowercase
    row_idx, col_idxs = letters_idx[0], letters_idx[1:]
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -