Why do you want a BigFloat matrix?
I am not saying it is not usefull, but in the examples here, you do the
calculation first and then the conversion. Also when dealing with BigFloat
it is important that you do not initialize them with with Float64 values
that are not exactly what you want them to be, because lots of base 10
fractions are not represetable as a base 2 fraction.
julia> BigFloat(0.1)
1.000000000000000055511151231257827021181583404541015625e-01 with 256 bits
of precision
julia> BigFloat("0.1")
1.000000000000000000000000000000000000000000000000000000000000000000000000000002e-01
with 256 bits of precision
The better option is to initialize with a string.
kl. 10:04:54 UTC+1 lørdag 4. januar 2014 skrev Andreas Noack Jensen
følgende:
>
> convert(Matrix{BigFloat},cor(A))
>
>
> 2014/1/4 paul analyst <[email protected] <javascript:>>
>
>> How to create a correlation matrix (each other) with precision Big Float
>>
>> r = BigFloat (cor (A))
>>
>> does not work ...
>>
>> Paul
>>
>
>
>
> --
> Med venlig hilsen
>
> Andreas Noack Jensen
>
