array | avg (const array &input, int dim=-1) |
average along a dimension | |
template<typename ty > | |
ty | avg (const array &input) |
average of all elements in array | |
array | var (const array &input, bool isbiased=false, int dim=-1) |
variance along a dimension | |
array | var (const array &input, const array weights, int dim=-1) |
weighted variance along a dimension | |
template<typename ty > | |
ty | var (const array &input, bool isbiased=false) |
variance of all elements in array | |
template<typename ty > | |
ty | var (const array &input, const array weights) |
weighted variance of all elements in array | |
array | cov (const array &input, bool isbiased=false) |
covariance of row observations | |
array | cov (const array &X, const array &Y, bool isbiased=false) |
covariance between two jointly distributed random variables | |
array | std (const array &in, bool isbiased=false, int dim=-1) |
standard deviation along a dimension. | |
template<typename ty > | |
ty | std (const array &input, bool isbiased=false) |
standard deviation of all elements in a vector. | |
array | median (const array &input, int dim=-1) |
median along a dimension | |
template<typename ty > | |
ty | median (const array &input) |
median of all elements in array | |
template<typename ty > | |
ty | corrcoef (const array &x, const array &y) |
Correlation coefficient between vectors x and y . |
array af::avg | ( | const array & | input, |
int | dim = -1 |
||
) |
average along a dimension
[in] | input | |
[in] | dim | dimension along which to operate (-1 indicates first nonsingleton dimension) |
dim
in the input ty af::avg | ( | const array & | input | ) |
average of all elements in array
[in] | input |
array af::var | ( | const array & | input, |
bool | isbiased = false , |
||
int | dim = -1 |
||
) |
variance along a dimension
[in] | input | |
[in] | isbiased | if true then use sample variance (N ); otherwise if false (default) use population variance (N - 1 ) where N=input.elements() |
[in] | dim | dimension along which to operate (-1 indicates first nonsingleton dimension) |
array af::var | ( | const array & | input, |
const array | weights, | ||
int | dim = -1 |
||
) |
weighted variance along a dimension
[in] | input | |
[in] | weights | placed on each element (same size as input ) |
[in] | dim | dimension along which to operate (-1 indicates first nonsingleton dimension) |
weights
instead of unit weight ty af::var | ( | const array & | input, |
bool | isbiased = false |
||
) |
variance of all elements in array
[in] | input | |
[in] | isbiased | if true then use sample variance (N ); otherwise if false (default) use population variance (N - 1 ) where N=input.elements() |
ty af::var | ( | const array & | input, |
const array | weights | ||
) |
weighted variance of all elements in array
[in] | input | |
[in] | weights | placed on each element (same size as input ) |
array af::cov | ( | const array & | input, |
bool | isbiased = false |
||
) |
covariance of row observations
[in] | input | rows are observations, columns are variables |
[in] | isbiased | if true then use sample variance (N ); otherwise if false (default) use population variance (N - 1 ) where N=input.elements() |
array af::cov | ( | const array & | X, |
const array & | Y, | ||
bool | isbiased = false |
||
) |
covariance between two jointly distributed random variables
[in] | X | |
[in] | Y | |
[in] | isbiased | if true then use sample variance (N ); otherwise if false (default) use population variance (N - 1 ) where N=input.elements() |
array af::std | ( | const array & | in, |
bool | isbiased = false , |
||
int | dim = -1 |
||
) |
standard deviation along a dimension.
[in] | in | The input matrix |
[in] | isbiased | if true then use sample variance (N ); otherwise if false (default) use population variance (N - 1 ) where N=input.elements() |
[in] | dim | dimension along which to operate (-1 indicates first nonsingleton dimension) |
dim
in the input ty af::std | ( | const array & | input, |
bool | isbiased = false |
||
) |
standard deviation of all elements in a vector.
[in] | input | |
[in] | isbiased | if true then use sample variance (N ); otherwise if false (default) use population variance (N - 1 ) where N=input.elements() |
array af::median | ( | const array & | input, |
int | dim = -1 |
||
) |
median along a dimension
[in] | input | |
[in] | dim | dimension along which to operate (-1 indicates first nonsingleton dimension) |
dim
in the input ty af::median | ( | const array & | input | ) |
median of all elements in array
[in] | input |
ty af::corrcoef | ( | const array & | x, |
const array & | y | ||
) |
Correlation coefficient between vectors x
and y
.