Mailing-List: contact user-help@mahout.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@mahout.apache.org
Received-SPF: pass (nike.apache.org: local policy includes SPF record at
 spf.trusted-forwarder.org)
Content-Type: text/plain; charset=windows-1252
Mime-Version: 1.0 (Mac OS X Mail 7.1 \(1827\))
Subject: Re: Popularity of recommender items
From: Pat Ferrel <pat@occamsmachete.com>
In-Reply-To: <C2955B0B-4147-4DE1-8B2D-98EAB832A6A2@gmail.com>
Date: Sat, 8 Feb 2014 08:50:10 -0800
Content-Transfer-Encoding: quoted-printable
Message-Id: <C11E4402-83FC-460B-B8AF-22EB72CA1446@gmail.com>
References: <F0157953-1514-4CD9-A758-B22D675FA31C@gmail.com>
 <CAH9ofMYXi0ZVgKHtqZOfiOOqLa-XvFQPy_oDDkqS15WBtUfWqw@mail.gmail.com>
 <CAEccTywM6+eKwBZB2kXFony4CR==OWLvR-J48+04hfNZz3OGnQ@mail.gmail.com>
 <CAJwFCa34M5jXFzbMCE-nVxe0oD9OiFvFYt=N7xG4VKqLQxzEzg@mail.gmail.com>
 <50977135-E391-41BB-9148-07E7E88B0409@gmail.com>
 <CAEccTywgV=GL1ik-L+3pxBiu6+oMNCPY+=_Rm7f3wgEG8zTHQQ@mail.gmail.com>
 <CAJwFCa0rx7ew=zPvWVQsS3m0jGzkWMcfsSO8H1F2wniLDu1gGw@mail.gmail.com>
 <25702BCB-0410-40B8-A786-4216436A0422@gmail.com>
 <C2955B0B-4147-4DE1-8B2D-98EAB832A6A2@gmail.com>
To: "user@mahout.apache.org" <user@mahout.apache.org>

Didn=92t mean to imply I had historical view data=97yet.

The Thompson sampling =91trick=92 looks useful for auto converging to =
the best of A/B versions and a replacement for dithering. Below you are =
proposing another case to replace dithering=97this time on a list of =
popular items? Dithering works on anything you can rank but Thompson =
Sampling usually implies a time dimension. The initial guess, first =
Thompson sample, could be thought of as a form of dithering I suppose? =
Haven=92t looked at the math but it wouldn=92t surprise me to find they =
are very similar things.

While we are talking about it, why aren=92t we adding things like =
cross-reccomendations, dithering, popularity, and other generally useful =
techniques into the Mahout recommenders? All the data is there to do =
these things, and they could be packaged in the same Mahout Jobs. They =
seem to be languishing a bit while technology and the art of =
recommendations moves on.

If we add temporal data to preference data a bunch of new features come =
to mind, like hot lists or asymmetric train/query preference history.

On Feb 6, 2014, at 9:43 PM, Ted Dunning <ted.dunning@gmail.com> wrote:

One way to deal with that is to build a model that predicts the ultimate =
number of views/plays/purchases for the item based on history so far. =20=


If this model can be made Bayesian enough to sample from the posterior =
distribution of total popularity, then you can use the Thomson sampling =
trick and sort by sampled total views rather than estimated total views. =
 That will give uncertain items (typically new ones) a chance to be =
shown in the ratings without flooding the list with newcomers. =20

Sent from my iPhone

> On Feb 7, 2014, at 3:38, Pat Ferrel <pat@occamsmachete.com> wrote:
>=20
> The particular thing I=92m looking at now is how to rank a list of =
items by some measure of popularity when you don=92t have a velocity. =
There is an introduction date though so another way to look at =
popularity might be to decay it with something like e^-t where t is it=92s=
 age. You can see the decay in the views histogram