SVD++ when there are no ratings

[Mario Levitin <ma...@gmail.com>]

Hi,
In my dataset there are no ratings. There is information only about which
users rented which movies.
I think SVD is not suitable in this case. What about SVD++? I heard that it
can be used with this kind of data. Are there any other alternatives (other
than neighborhood based ones)?

Thanks
mario

Re: SVD++ when there are no ratings

[Mario Levitin <ma...@gmail.com>]

Thanks for your reply Sebastian,
One more question: What about SVD++, isn't it different from using
ALSWRFactorizer
with usesImplicitFeedback = true.


On Thu, Jan 9, 2014 at 3:04 PM, Sebastian Schelter <ss...@apache.org> wrote:

> I would recommend to try an item-based recommender with loglikelihood
> ratio as similarity measure or an SVDRecommender with an ALSWRFactorizer
> with usesImplicitFeedback = true
>
> --sebastian
>
>
> On 01/09/2014 01:57 PM, Mario Levitin wrote:
>
>> Hi,
>> In my dataset there are no ratings. There is information only about which
>> users rented which movies.
>> I think SVD is not suitable in this case. What about SVD++? I heard that
>> it
>> can be used with this kind of data. Are there any other alternatives
>> (other
>> than neighborhood based ones)?
>>
>> Thanks
>> mario
>>
>>
>

Re: SVD++ when there are no ratings

[Sebastian Schelter <ss...@apache.org>]

I would recommend to try an item-based recommender with loglikelihood 
ratio as similarity measure or an SVDRecommender with an ALSWRFactorizer 
with usesImplicitFeedback = true

--sebastian

On 01/09/2014 01:57 PM, Mario Levitin wrote:
> Hi,
> In my dataset there are no ratings. There is information only about which
> users rented which movies.
> I think SVD is not suitable in this case. What about SVD++? I heard that it
> can be used with this kind of data. Are there any other alternatives (other
> than neighborhood based ones)?
>
> Thanks
> mario
>