小编man*_*pig的帖子

我已经在Python中实现了Perceptron学习算法,如下所示.即使有500,000次迭代,它仍然不会收敛.

我有一个带有目标矢量Y的训练数据矩阵X和一个要优化的权重矢量w.

我的更新规则是:

while(exist_mistakes): 
    # dot product to check for mistakes
    output = [np.sign(np.dot(X[i], w)) == Y[i] for i in range(0, len(X))]

    # find index of mistake. (choose randomly in order to avoid repeating same index.) 
    n = random.randint(0, len(X)-1)
    while(output[n]): # if output is true here, choose again
        n = random.randint(0, len(X)-1)

    # once we have found a mistake, update
    w = w + Y[n]*X[n] 

这是错的吗？或者为什么即使在500,000次迭代后它也没有收敛？

我正在使用 BigQuery 运行带有多个WITH 子句的查询，如下所示：

WITH subq_L1 AS (
    SELECT
        id,
        metric,
        COUNT(DISTINCT IF(DATEDIFF(TIMESTAMP('{execution_date:s}'), TIMESTAMP(ds)) < 1, ds, NULL)) AS L1
    FROM
        [xxx:yyy.zzz]
    GROUP EACH BY
        id, metric
),
subq_L7 AS (
    SELECT
        id,
        metric,
        COUNT(DISTINCT IF(DATEDIFF(TIMESTAMP('{execution_date:s}'), TIMESTAMP(ds)) < 7, ds, NULL)) AS L7
    FROM
        [xxx:yyy.zzz]
    GROUP EACH BY
        id, metric
),
subq_L14 AS (
    SELECT
        id,
        metric,
        COUNT(DISTINCT IF(DATEDIFF(TIMESTAMP('{execution_date:s}'), TIMESTAMP(ds)) < 14, ds, NULL)) AS L14
    FROM
        [xxx:yyy.zzz]
    GROUP EACH BY
        id, metric
)
SELECT * FROM subq_L1
UNION ALL …