猜你喜欢(Guess what you like)
under review
Simen
为碎片信息进行主题推荐呢?比如用户设定一个主题描述,然后以后的碎片信息,推荐与哪些主题,相关,让用户在这个基础上,选择,并关联。感觉这个还是NLP的模型
1
1H
Simen: 这个提议不错,可以先做个简单的TF-IDF关键词提取,然后基于关键词也就是你说的主题,来推荐。
Boninall | Meogic
under review
Simen
内核是智能推荐算法?
1
1H
Simen: 没那么高深,仅仅只是属性的相似度。推荐算法有人和物的交互,离线本地软件只有一个人。。。就用不了协同过滤了
1
1H
The recommendation of similar objects is done based on the similarity between object attributes.
The objects here are like, documents and each document inside Zotero.
The attributes here are
Attributes set by the user to the document or other objects on the edit page
Attributes given by the act of double chaining
For example, if node A of a mind map has a document double-linked to its child node B, then the document is given the attribute "node A->node B".