《推荐系统实践》读后的一些想法
- 说在前头的话: 在对这本书进行阅读之前,我对于大模块知识的学习方式一无所知,所以就想借由学习推荐系统的知识,来构建属于自己的大模块知识学习方式。我把这个方向上的学习,更倾向于一种对于学习本身的探索,而不强求自己真正能掌握多少知识。出于以上的原因,我将要记录的东西更偏向于一种阅读完之后的自己的理解,相当于把别人的东西吃进去,嚼两口再吐出来。其中所产生的信息缺失或信息差错,还望读者体谅,对于一些存疑的地方,翻阅原著是最佳的方式,若是有自己的思考,敬请斧正。
- 所以说,如果想对这方面有所了解的话,那我推荐您阅读原著,然后把文档中内容当成是复习或者是对照学习的资料,文档中的内容缺少了原文中生动形象的例子,对于刚接触者来说可能难以理解
- 小人不才,在专业知识匮乏和自身水平较低的情况下,擅自对这书的内容下一个定义。全书所围绕的主题重点在于从数据的角度出发,进行推荐系统的构造,重点着墨于包括:基于用户的行为数据,基于标签数据,基于上下文信息,给予社交网络数据,这些一早就写在目录中的数据。
- 书中的讲述深入浅出,如果愿意用心阅读作者的描述的话,很多的晦涩概念都被作者准确的描述了出来,再辅以一些实际例子,让整本书的阅读难度甚至降低到了兴趣者能读懂的程度。足以见到作者高超的笔力和深厚的理解。
- 当然书中并没有一个完整推荐系统的代码,这方面的空缺建议辅佐另一本《深度学习推荐系统》和 fun-rec 开源社区中的代码进行学习。
- 对于有想法了解推荐系统的人来说,这本书可以是一本合适的入门作品,书中对于推荐系统的定义和描述,足以让有想法的人对于这个领域有一个大概的了解。