Kabbur等提出了进一步表达用户信息的因子项目相似性模型factored item similarity modelsFISM。此模型本质上是基于物品的协同推荐算法它将用户历史评分过的物品作为特征属性来得到用户表示用户表示与物品表示的内积来表达用户对物品的偏好。SVD++模型则结合了基于用户的推荐算法和基于物品的推荐算法这两者的优势。重写一下上面这段话

Kabbur and others proposed a Factored Item Similarity Model (FISM) that further expresses user information through factor items. This model is essentially based on item collaborative recommendation algorithms, using items that the user has historically rated as feature attributes to obtain user representations. The user's preference for an item is expressed by the inner product of the user representation and the item representation. The SVD++ model combines the strengths of both user-based and item-based recommendation algorithms.