On 13 Oct 2011, at 22:13, Ted Dunning <[email protected]> wrote:
> I think adjoining and adding are the only two ops necessary.
That sounds great. Plus usual args eg --overwrite, --help. For weights, from my
side it'd be premature to request anything but I'd guess I could do most of
what I need by interfering with each matrix earlier in the pipeline...
Dan
> On Thu, Oct 13, 2011 at 8:09 PM, Lance Norskog <[email protected]> wrote:
>
>> A bin/mahout job to to turn A and B into rows of a1,a2,...aN,b1,b2...bN?
>> What extra options would you like? For example, would you want to apply
>> different weights to matrix A v.s. matrix B?
>>
>> On Thu, Oct 13, 2011 at 10:11 AM, Ted Dunning <[email protected]>
>> wrote:
>>
>>> This is relatively easy to do at the code level, but I don't know of a
>>> command line level way to do this. As you suggest this involves
>> adjoining
>>> the two matrices.
>>>
>>> If you use the feature hashing, adjoining works, but it is also possible
>> to
>>> simply add the two matrices (assuming conformal sizes).
>>>
>>> On Thu, Oct 13, 2011 at 4:55 PM, Dan Brickley <[email protected]> wrote:
>>>
>>>> I have a matrix of 100,000 items x 30k features; and another of those
>>>> same 100,000 items, x however-many different features (from n-gram
>>>> collocation extraction). In current app, these are library holdings
>>>> and subject codes + extracted phrases. (later these should be 14
>>>> million items by somewhat but not shockingly larger feature space, if
>>>> that is useful to know)
>>>>
>>>> I'd like to compose these into a larger unified feature matrix, with
>>>> same row structure, and with feature columns drawing from both input
>>>> matrices. So far in this work I've managed to get by using bin/mahout
>>>> rather than firing up Eclipse and messing with Java; I'd be happy to
>>>> learn I can continue in this work style. But if custom code is needed
>>>> that's fine. Either way, some pointer would be much appreciated...
>>>>
>>>> thanks,
>>>>
>>>> Dan
>>>>
>>>
>>
>>
>>
>> --
>> Lance Norskog
>> [email protected]
>>
