Show HN: Osmoar: A Bayesian Suggestion Engine | Hacker News Reader

Show HN: Osmoar: A Bayesian Suggestion Engine

osmoar.com

6 pointszacstewart6 comments

A friend and I are launching an alpha first thing in 2011. We're trying to gain some momentum and whatnot, so we're opening early registration as of tonight.

We're planning to develop a Bayesian network to derive suggestions. If you're interested in seeing our progress, check it out.

Edit: Any and all suggestions, criticism, advice, etc is highly appreciated!

Open articleView on HN

why three personal items?

Also, more info would be nice.

My assumption is that they will suggest to you other possessions that you might want to buy, based on the three favorites you provide.

Good assumption, but we're doing more with it than that. There's also an "activities as recipes of items" aspect we're working on.

We don't want to just push more crap from amazon on users. We want to help users become aware of activities they could be doing with the things they own as well.

"789 users that own a MacBook and a DSL camera are doing photography. Maybe you'd like to try it"

Try using a camera for photography?! I might need a better example to see the value proposition :o)

Haha. Sorry. Pretty much all your activities can be broken down into a recipe of the things you do them with, the place you do them, and a description of what it is you do.

It will help you find something to do with your possessions in the place you live. E.g., if you have a bike, it might suggest you right critical mass on the last Friday of the month. If you have a dog attend the people with dogs party at your local park this weekend. I can't really come up with a great example myself, but once it's in action, I can see some pretty interesting "recipes" of place, items, people and activities arising.

Value insofar as monetization, is obviously affiliate income from retailers. We're planning to establish an API for any retailer that wants to sign up to add their link to an item's page, which is also helpful to users so they can sort by price and compare retailers.

It's basically so you have a profile started when we launch. We're planning on users adding a whole inventory of their things via iPhone barcode scanner, ISBN, search, etc.

From that inventory, we calculate interests in other items. From aggregate user input, we suggest interesting things to do with things you have, things you might want, etc.

I'm still learning how Bayesian networks work, and we might decide on another algorithm entirely, but we'll see how it goes :)

Thanks for your suggestion. We're building an FAQ out of questions like the one you just asked! Is there anything I could explain better?

Reply on news.ycombinator.com