设计问题:搜索产品还是自动贴标产品? (搜索与机器学习)
我正在开展一个项目,该项目有许多零售商独立上传他们的产品信息,换句话说,完全相同的产品可能有不同的名称。例如,new ipad,或 ipad 或 Great ipad 等...
我想向用户呈现一个直观的界面,以便当他搜索 ipad 或多个 ipad 时,他会得到一个 ipad 作为搜索结果,并并列所有价格。或者可能有 2 个搜索结果,第一个搜索结果是旧 ipad,包含不同零售商的所有价格,另一个搜索结果是 ipad 2,并且价格一个挨着一个。
具体例子: 数据库包含:new ipad 1、ipad 1、great ipad 1、new ipad 2、ipad 2、great ipad 2 当用户搜索 ipad 时,他得到:
- ipad1:零售商 1 的价格 = 100$,零售商 2 的价格 = 120$ 等...
- ipad2:零售商 2 的价格 = 100$,零售商 2 的价格 = 120$ 等... ?
这主要是一个搜索特定问题吗 或者机器学习、自动标记?
有谁知道 Rails3 中已经实现了类似的东西吗?
谢谢
I'm working on a project, that has many retailers uploading their products information independently, in other words the exact same product might have various name. For instance, new ipad, or ipad or Great ipad etc...
I want to present to the user an intuitive interface so that when he searches for ipad or ipads, he gets as search result one ipad with all the prices juxtaposed. Or maybe 2 search results, the first one with the old ipad with all the prices from different retailers, and another search result with the ipad 2 and also with the prices one next to the other.
Concrete example:
database contains: new ipad 1, ipad 1, great ipad 1, new ipad 2, ipad 2, great ipad 2
when user searches for ipad, he gets:
- ipad1: price of retailer 1 = 100$, price of retailer 2 = 120$ etc...
- ipad2: price of retailer 2 = 100$, price of retailer 2 = 120$ etc...
Is this mainly a search specific problem? OR machine learning, autolabeling?
And does anyone know something that has been implemented like that in Rails3?
Thanks
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(1)
这是重复删除问题。地图、目录正在处理它。这个问题的解决方案有很多种,但成功程度各不相同。
最简单的一个是识别同义词和无用词(很棒,有趣)等,并使用这些信息进行项目匹配。一些不清楚的案例您应该发送给人工审核。
示例“new Apple Ipad”与“Great ipad”=删除无用词=> “Apple ipad”与“ipad”=通过子字符串相等/不明确=>平等的
It is duplicate removing problem. Maps, catalogs are dealing with it. The are many solutions to this problem with different success degree.
Most simple one is to identify synonyms and useless words(Great, fun) etc and do items matching using this information. Some unclear case you should send to human for review.
Example "new Apple Ipad" vs "Great ipad" =remove usless word=> "Apple ipad" vs "ipad" =equal by substring/to ambiguous=> equal