序列化推荐
什么是序列化推荐?
在现实世界中,用户的前后行为都存在强烈的关联性与因果性,将用户过去的这一系列交互行为视作用户行为序列 u={
传统的推荐系统,例如基于内容和协同过滤的推荐系统,以一种静态的方式建模用户和商品的交互并且只可以捕获用户广义的喜好。
而 SRSs 则是将用户和商品的交互建模为一个动态的序列并且利用序列的依赖性来活捉当前和最近用户的喜好。
在现实世界中,用户的前后行为都存在强烈的关联性与因果性,将用户过去的这一系列交互行为视作用户行为序列 u={
传统的推荐系统,例如基于内容和协同过滤的推荐系统,以一种静态的方式建模用户和商品的交互并且只可以捕获用户广义的喜好。
而 SRSs 则是将用户和商品的交互建模为一个动态的序列并且利用序列的依赖性来活捉当前和最近用户的喜好。