跳到主要内容

类别与离散均匀

整数槽位是 UBIGINT/BIGINT 参数;非整数字面量找不到匹配的签名。

Categorical(类别)​

参数:probabilities —— 未归一化的概率 LIST(DOUBLE);statrs 内部会自动归一。

sr_categorical_pmf(x, probabilities)​

签名:sr_categorical_pmf(x UBIGINT, probabilities DOUBLE[]) -> DOUBLE

-- [1, 2, 1] 归一化为 [0.25, 0.5, 0.25];pmf(1) = 0.5
SELECT sr_categorical_pmf(1, [1.0, 2.0, 1.0])
-- 0.5

sr_categorical_ln_pmf(x, probabilities)​

签名:sr_categorical_ln_pmf(x UBIGINT, probabilities DOUBLE[]) -> DOUBLE

SELECT sr_categorical_ln_pmf(1, [1.0, 2.0, 1.0])

sr_categorical_cdf(x, probabilities)​

签名:sr_categorical_cdf(x UBIGINT, probabilities DOUBLE[]) -> DOUBLE

SELECT sr_categorical_cdf(1, [1.0, 2.0, 1.0])
-- 0.75

sr_categorical_sf(x, probabilities)​

签名:sr_categorical_sf(x UBIGINT, probabilities DOUBLE[]) -> DOUBLE

SELECT sr_categorical_sf(1, [1.0, 2.0, 1.0])

sr_categorical_quantile(p, probabilities)​

签名:sr_categorical_quantile(p DOUBLE, probabilities DOUBLE[]) -> DOUBLE

SELECT sr_categorical_quantile(0.5, [1.0, 2.0, 1.0])
-- 1.0

pmf 中 x 不是整数:

SELECT sr_categorical_pmf(1.5, [1.0, 2.0, 1.0])
-- error: no matching signature: x is UBIGINT, so 1.5 does not bind

sr_categorical_entropy(probs)​

签名:sr_categorical_entropy(probs DOUBLE[]) -> DOUBLE

Categorical(类别)的熵。

SELECT sr_categorical_entropy([1.0, 2.0, 1.0])
-- 1.0397207708399179

sr_categorical_max(probs)​

签名:sr_categorical_max(probs DOUBLE[]) -> DOUBLE

Categorical(类别)的最大值。

SELECT sr_categorical_max([1.0, 2.0, 1.0])
-- 2.0

sr_categorical_mean(probs)​

签名:sr_categorical_mean(probs DOUBLE[]) -> DOUBLE

Categorical(类别)的均值。

SELECT sr_categorical_mean([1.0, 2.0, 1.0])
-- 1.0

sr_categorical_median(probs)​

签名:sr_categorical_median(probs DOUBLE[]) -> DOUBLE

Categorical(类别)的中位数。

SELECT sr_categorical_median([1.0, 2.0, 1.0])
-- 1.0

sr_categorical_min(probs)​

签名:sr_categorical_min(probs DOUBLE[]) -> DOUBLE

Categorical(类别)的最小值。

SELECT sr_categorical_min([1.0, 2.0, 1.0])
-- 0.0

sr_categorical_skewness(probs)​

签名:sr_categorical_skewness(probs DOUBLE[]) -> DOUBLE

Categorical(类别)的偏度。

SELECT sr_categorical_skewness([1.0, 2.0, 1.0])
-- NULL

sr_categorical_std_dev(probs)​

签名:sr_categorical_std_dev(probs DOUBLE[]) -> DOUBLE

Categorical(类别)的标准差。

SELECT sr_categorical_std_dev([1.0, 2.0, 1.0])
-- 0.7071067811865476

sr_categorical_variance(probs)​

签名:sr_categorical_variance(probs DOUBLE[]) -> DOUBLE

Categorical(类别)的方差。

SELECT sr_categorical_variance([1.0, 2.0, 1.0])
-- 0.5

Discrete uniform(离散均匀)​

参数:min、max——BIGINT。支撑:[min, max] 内的整数。

sr_discrete_uniform_pmf(x, min, max)​

签名:sr_discrete_uniform_pmf(x BIGINT, min BIGINT, max BIGINT) -> DOUBLE

SELECT sr_discrete_uniform_pmf(2, 1, 6)
-- 0.16666666666666666

sr_discrete_uniform_ln_pmf(x, min, max)​

签名:sr_discrete_uniform_ln_pmf(x BIGINT, min BIGINT, max BIGINT) -> DOUBLE

SELECT sr_discrete_uniform_ln_pmf(2, 1, 6)

sr_discrete_uniform_cdf(x, min, max)​

签名:sr_discrete_uniform_cdf(x BIGINT, min BIGINT, max BIGINT) -> DOUBLE

SELECT sr_discrete_uniform_cdf(3, 1, 6)

sr_discrete_uniform_sf(x, min, max)​

签名:sr_discrete_uniform_sf(x BIGINT, min BIGINT, max BIGINT) -> DOUBLE

SELECT sr_discrete_uniform_sf(3, 1, 6)

sr_discrete_uniform_quantile(p, min, max)​

签名:sr_discrete_uniform_quantile(p DOUBLE, min BIGINT, max BIGINT) -> DOUBLE

SELECT sr_discrete_uniform_quantile(0.5, 1, 6)
-- 3.0

sr_discrete_uniform_entropy(min, max)​

签名:sr_discrete_uniform_entropy(min BIGINT, max BIGINT) -> DOUBLE

Discrete uniform(离散均匀)的熵。

SELECT sr_discrete_uniform_entropy(1, 6)
-- 1.791759469228055

sr_discrete_uniform_max(min, max)​

签名:sr_discrete_uniform_max(min BIGINT, max BIGINT) -> DOUBLE

Discrete uniform(离散均匀)的最大值。

SELECT sr_discrete_uniform_max(1, 6)
-- 6.0

sr_discrete_uniform_mean(min, max)​

签名:sr_discrete_uniform_mean(min BIGINT, max BIGINT) -> DOUBLE

Discrete uniform(离散均匀)的均值。

SELECT sr_discrete_uniform_mean(1, 6)
-- 3.5

sr_discrete_uniform_median(min, max)​

签名:sr_discrete_uniform_median(min BIGINT, max BIGINT) -> DOUBLE

Discrete uniform(离散均匀)的中位数。

SELECT sr_discrete_uniform_median(1, 6)
-- 3.5

sr_discrete_uniform_min(min, max)​

签名:sr_discrete_uniform_min(min BIGINT, max BIGINT) -> DOUBLE

Discrete uniform(离散均匀)的最小值。

SELECT sr_discrete_uniform_min(1, 6)
-- 1.0

sr_discrete_uniform_mode(min, max)​

签名:sr_discrete_uniform_mode(min BIGINT, max BIGINT) -> DOUBLE

Discrete uniform(离散均匀)的众数。

SELECT sr_discrete_uniform_mode(1, 6)
-- 3.0

sr_discrete_uniform_skewness(min, max)​

签名:sr_discrete_uniform_skewness(min BIGINT, max BIGINT) -> DOUBLE

Discrete uniform(离散均匀)的偏度。

SELECT sr_discrete_uniform_skewness(1, 6)
-- 0.0

sr_discrete_uniform_std_dev(min, max)​

签名:sr_discrete_uniform_std_dev(min BIGINT, max BIGINT) -> DOUBLE

Discrete uniform(离散均匀)的标准差。

SELECT sr_discrete_uniform_std_dev(1, 6)
-- 1.707825127659933

sr_discrete_uniform_variance(min, max)​

签名:sr_discrete_uniform_variance(min BIGINT, max BIGINT) -> DOUBLE

Discrete uniform(离散均匀)的方差。

SELECT sr_discrete_uniform_variance(1, 6)
-- 2.9166666666666665