參考:http://scikit-learn.org/stable/modules/scaling_strategies.html
对于examples、features(或者两者)数量非常大的情况,挑战传统的方法要解决两个问题:内存和效率。办法是Out-of-core (or “external memory”) learning。
有三种方法能够实现out-of-core。各自是:
1、Streaming instances(流体化实例):
简单说就是。instances是一个一个来的。详细实现不在scikit-learn文档范围。
2、Extracting features:
简单说就是利用different feature extraction methods(翻译之后的文章:http://blog.csdn.net/mmc2015/article/details/46992105)实现大数据提取实用数据。简化内存、提高效率。不细讲。
3、Incremental learning:
all
estimators implementing the partial_fit API
are candidates。
the
ability to learn incrementally from a mini-batch of instances (sometimes called “online learning”) is key to out-of-core learning as it guarantees that at any given time there will be only a small amount of instances in the main memory。
全部实现 partial_fit API 的estimators都能够实现增量学习,包含:
-
- Clustering
-
- Decomposition / feature Extraction
注意:对于分类问题,因为incremental learner可能不知道全部的classes有哪些,所以第一次调用partial_fit时,最好人工设定參数 classes= ,指明全部类别。
4、Examples:
a example of Out-of-core classification of text documents. 通过样例能够更好理解上面的内容。