是的。并不是所有的Mahout都依赖于Hadoop，尽管很多。如果您使用依赖于Hadoop的作品，当然，您需要Hadoop。但是，例如，有一个实用的推荐引擎代码库，它不使用Hadoop。

Yes. Not all of Mahout depends on Hadoop, though much does. If you use a piece that depends on Hadoop, of course, you need Hadoop. But for example there is a substantial recommender engine code base that does not use Hadoop.

您可以在Java程序中嵌入本地Hadoop集群/ worker。

You can embed a local Hadoop cluster/worker in a Java program.