利用Map/Reduce的机器学习算法库 Apache Mahout 0.6 发布

     <p><a href="/misc/goto?guid=4958328385917577165" target="_blank">Mahout</a>是一个利用Map/Reduce的机器学习算法库，其思想源于斯坦福大学几个学者在2006年的nips会议上发表的一篇文章“Map- Reduct for Machine Learning on Multicore"</p>    <p>Apache Mahout 0.6 发布了，建议所有开发者升级，该版本主要改进包括：</p>    <ul>     <li>Improved Decision Tree performance and added support for regression problems</li>     <li>New LDA implementation using Collapsed Variational Bayes 0th Derivative Approximation</li>     <li>Reduced runtime of LanczosSolver tests</li>     <li>K-Trusses, Top-Down and Bottom-Up clustering, Random Walk with Restarts implementation</li>     <li>Reduced runtime of dot product between vectors</li>     <li>Added MongoDB and Cassandra DataModel support</li>     <li>Increased efficiency of parallel ALS matrix factorization</li>     <li>SSVD enhancements</li>     <li>Performance improvements in RowSimilarityJob, TransposeJob</li>     <li>Added numerous clustering display examples</li>     <li>Many bug fixes, refactorings, and other small improvements</li>    </ul>    <p>完整列表请看：<a href="https://issues.apache.org/jira/secure/ReleaseNote.jspa?projectId=12310751&version=12316364">release notes</a>.</p>    <p>下载地址：<a href="/misc/goto?guid=4958328387526338701">Apache mirrors</a>.</p>