Both algorithms execute in worst case O(log n) iterations and are
quite general. They would likely be able to server as a basis for more
complex algorithms.
I will have to take a closer look at the Mahout data structures before
I can be certain how hard it would be. However, the connectivity
algorithm should be quite trivial as it is only a few lines of code.
The MST algorithm may not fit as nicely but I will do a little
research and get back to you.
Perhaps to begin with we can work on integrating the graph
connectivity algorithm. This should allow me to become more familiar
with the Mahout code base and its developers before tackling more
complex algorithms.
Thanks,
Neal.
On Mon, May 31, 2010 at 2:00 PM, Ted Dunning <ted.dunning@gmail.com> wrote:
> These might be very interesting.
>
> How efficient are these algorithms? How general?
>
> How hard would it be to merge these into Mahout style matrix and vector data
> structures?
>
> On Mon, May 31, 2010 at 1:48 PM, Neal Clark <nclark.uvic@gmail.com> wrote:
>
>> As part of my research for my masters thesis I have developed a number
>> of distributed MapReudce graph algorithms.
>>
>
