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: domain of timsmith_s@hotmail.com
 designates 65.55.116.47 as permitted sender)
Message-ID: <BLU173-W288969C3ACCCA9EC275BACF8B20@phx.gbl>
Content-Type: multipart/alternative;
	boundary="_6a79a8b4-4a75-4927-999f-de95447862ad_"
From: Tim Smith <timsmith_s@hotmail.com>
To: "user@mahout.apache.org" <user@mahout.apache.org>
Subject: RE: Item recommendation w/o users or preferences
Date: Fri, 10 Jan 2014 22:01:00 -0600
Importance: Normal
In-Reply-To: 
 <4B7FE316905F4B43A5718F1AE91BFB348EBAAEF0@PHCMPR1G.safeway01.ad.safeway.com>
References: 
 <BLU173-W48EE76CA15A2065FD45E52F8B20@phx.gbl>,<4B7FE316905F4B43A5718F1AE91BFB348EBAAEF0@PHCMPR1G.safeway01.ad.safeway.com>
MIME-Version: 1.0

--_6a79a8b4-4a75-4927-999f-de95447862ad_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Excellent question.  Given who you work for=2C just assume a customer comes=
 into a retail location and goes to pay at the checkout.  They do not ident=
ity themselves (no loyalty/club card) and use cash (trying to make the poin=
t that we have no idea who this consumer is right at this moment=2C and may=
 never will).  So rather than having Catalina print out coupons after the f=
act=2C say I want to make an offer right there at the POS during their tran=
saction.  I realize that this is a bit problematic at a grocery store=2C bu=
t our scenario has a clerk behind a counter with these items close at hand.=
  So all I have is their current basket and the baskets of previous anonymo=
us purchases.  Clear?

> From: Rachel.Owsley@safeway.com
> To: user@mahout.apache.org
> Subject: RE: Item recommendation w/o users or preferences
> Date: Sat=2C 11 Jan 2014 03:49:53 +0000
>=20
> Hi Tim=2C=20
>=20
> By not having user or preference information=2C it's not clear to me-- do=
 you mean you have no demographic information=2C but you have email or some=
 IP address-- some way to track the user?
>=20
> It is possible to generate recommendations on purchase history=2C by look=
ing at the user's transactions and inferring a preference from what they bu=
y the most frequently. I used to work for a company that had transaction hi=
story=2C but it was anonymized-- all the user's activity was tied to an ano=
nymous token. They didn't even have the name or gender. If you know a custo=
mer's card #=2C you could relate the card #   as their "user_id" and use th=
e count or monetary value of their transactions for a specific item as a pr=
eference for that item. Try something like conditional probability-- the pr=
obability that you will buy one thing given that you bought another. By gen=
erating a set of pairs (item a being the user has bought=2C and item b bein=
g the one they have not purchased)=2C you can determine the probability tha=
t they will by item b=2C given that they bought item A.=20
>=20
> Still=2C if you know nothing about a person at all=2C and don't even have=
 a way to distinguish them on your website=2C then recommendation won't rea=
lly help much because how will you actually give the user recommendations? =
You could consider using market basket analysis to tell you what other item=
s a person might put in his/her cart. I've done market basket analysis befo=
re. It is necessary to do a lot of "pruning" with market basket analysis=2C=
 because a lot of the frequent pairs are not very useful. But through some =
careful analysis=2C you may find interesting combinations of items that wil=
l help your business in terms of cross selling/promotion. I am  looking at =
sequential basket analysis right now. If I buy items x1 through x4=2C what =
is the probability that a certain item will be the next one? You might be a=
ble to use something market basket (fpgrowth) or maybe a markov model to de=
termine the next item in sequence.=20
>=20
> Good luck with this. If you could share the type of data you do have avai=
lable=2C it would be helpful.
>=20
> Rachel
> ________________________________________
> From: Tim Smith [timsmith_s@hotmail.com]
> Sent: Friday=2C January 10=2C 2014 5:27 PM
> To: user@mahout.apache.org
> Subject: Item recommendation w/o users or preferences
>=20
> Say I have a retail organization that doesn't sell a diverse set of produ=
cts=2C eg 2000=2C but has many small transactions.  Also say that I don't h=
ave any user or preference information.  Is it reasonable to use pattern mi=
ning (market baskets) and recommend items based on a set of thresholds for =
support=2C confidence=2C and lift?  If not=2C what are my options?
>=20
> "Email Firewall" made the following annotations.
> -------------------------------------------------------------------------=
-----
>=20
> Warning:=20
> All e-mail sent to this address will be received by the corporate e-mail =
system=2C and is subject to archival and review by someone other than the r=
ecipient.  This e-mail may contain proprietary information and is intended =
only for the use of the intended recipient(s).  If the reader of this messa=
ge is not the intended recipient(s)=2C you are notified that you have recei=
ved this message in error and that any review=2C dissemination=2C distribut=
ion or copying of this message is strictly prohibited.  If you have receive=
d this message in error=2C please notify the sender immediately.  =20
> =20
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D
>=20
 		 	   		  =

--_6a79a8b4-4a75-4927-999f-de95447862ad_--