The UnivariateKDE object is just a Range and an Array of the same length, 
so you could also just save that information, and recreate the object via 
UnivariateKDE(range,density).

On Tuesday, 24 March 2015 13:32:46 UTC+1, René Donner wrote:
>
> You can try the jld* functions of the HDF5.jl package, they can save 
> custom types: https://github.com/timholy/HDF5.jl
>
> Am Dienstag, 24. März 2015 13:21:02 UTC+1 schrieb Christopher Fisher:
>>
>> I would like to save the output of a kernel density estimate function for 
>> future use. The KernelDensity package returns an object of the 
>> type UnivariateKDE{FloatRange{Float64}} (constructor with 1 method). 
>> Ideally, I would like to save upwards of 1000 of these for future use 
>> because the model I am simulating is computationally intensive. Any help 
>> would be greatly appreciated. 
>>
>

Reply via email to