> Another topic in the last chat was the broad subject of collaborative
> filtering (the technique that Amazon and Netflicks, for example, use
> for making recommendations).  I think that the core concept has
> applicability to making wine recommendations, but the current
> practices are not refined enough.

> As I see it, the major shortcoming we have is that our rating and
> scoring systems do not provide enough discrimination to provide the
> groupings necessary to effectively use the collaborative filtering
> algorithms.  My opinion, of course; I'm interested in discussing this
> topic and perhaps arriving at some methods for achieving that
> discrimination necessary (see the thread I started on "Calibrating
> Palates, for example)>

> Are any of the participants in this Wine 2.0 group using collaborative
> filtering?  If so, please describe your approach and how it is
> working.

> Ward Bell
> WineOnline Minnesota
> (A Pre-Alpha Development)

> Ward

> Snooth (www.snooth.com) is about to launch and thats exactly what
> we're doing. Take a look at the blog.

> I dont want to spill any details yet, but can talk to you once we open
> to the public.

> Best
> Philip

> On Apr 13, 11:11 am, "Ward" <wineonl...@gmail.com> wrote:

> > Another topic in the last chat was the broad subject of collaborative
> > filtering (the technique that Amazon and Netflicks, for example, use
> > for making recommendations).  I think that the core concept has
> > applicability to making wine recommendations, but the current
> > practices are not refined enough.

> > As I see it, the major shortcoming we have is that our rating and
> > scoring systems do not provide enough discrimination to provide the
> > groupings necessary to effectively use the collaborative filtering
> > algorithms.  My opinion, of course; I'm interested in discussing this
> > topic and perhaps arriving at some methods for achieving that
> > discrimination necessary (see the thread I started on "Calibrating
> > Palates, for example)>

> > Are any of the participants in this Wine 2.0 group using collaborative
> > filtering?  If so, please describe your approach and how it is
> > working.

> > Ward Bell
> > WineOnline Minnesota
> > (A Pre-Alpha Development)- Hide quoted text -

> - Show quoted text -

> Thanks, Philip, for your response and pointer to your blog.  It does
> sound like the program you are developing has some interesting ways of
> approaching the issues of building a powerful recommendation engine.
> I look forward to point in time when you move out of your alpha/beta
> mode and into full release.

> I spent some time going through all the posts you've made on your
> blog:  There are some interesting things that you have revealed even
> while being circumspect.  I'm impressed that you have raises some
> venture capital.

> Is there anything else you can say, perhaps in this somewhat limited
> venue (there are only a couple handsful of subscribers at this time)?

> And, are there others who might have information they are willing to
> share in this space?

> Ward Bell
> WineOnline Minnesota
> (Pre-Alpha Stage Development)

> On Apr 13, 11:41 am, "Philip James" <philip.ja...@gmail.com> wrote:

> > Ward

> > Snooth (www.snooth.com) is about to launch and thats exactly what
> > we're doing. Take a look at the blog.

> > I dont want to spill any details yet, but can talk to you once we open
> > to the public.

> > Best
> > Philip

> > On Apr 13, 11:11 am, "Ward" <wineonl...@gmail.com> wrote:

> > > Another topic in the last chat was the broad subject of collaborative
> > > filtering (the technique that Amazon and Netflicks, for example, use
> > > for making recommendations).  I think that the core concept has
> > > applicability to making wine recommendations, but the current
> > > practices are not refined enough.

> > > As I see it, the major shortcoming we have is that our rating and
> > > scoring systems do not provide enough discrimination to provide the
> > > groupings necessary to effectively use the collaborative filtering
> > > algorithms.  My opinion, of course; I'm interested in discussing this
> > > topic and perhaps arriving at some methods for achieving that
> > > discrimination necessary (see the thread I started on "Calibrating
> > > Palates, for example)>

> > > Are any of the participants in this Wine 2.0 group using collaborative
> > > filtering?  If so, please describe your approach and how it is
> > > working.

> > > Ward Bell
> > > WineOnline Minnesota
> > > (A Pre-Alpha Development)- Hide quoted text -

> > - Show quoted text -