英语翻译6.ConclusionWe have presented an innovative approach for

英语翻译
6.Conclusion
We have presented an innovative approach for effciently generating library recommendations based on user purchase histories with restricted random walks.The quality of the recommendations as well as the computational performance of the algorithm are promising.
For the next steps,we will make the interface public in order to gain user feedback on the recommendations and perform an evaluation based on the users’ judgments.As for the theoretical side of the algorithm,measures for fine tuning the parameters will be scrutinized.Another important aspect is the update of the recommendations in accordance to the changes of the similarity graph induced by the continued use of the library’s web interface.Each week,the graph is updated with the usage histories of the period.Given the small volume of these data sets in comparison to the overall size of the document data,it is highly desirable to have an update procedure that is able to effciently incorporate these changes into the clusters and thus into the recommendation lists.The investigation of the influence of the number of walks that are started from each node as well as a better understanding of the asymptotic behavior of the walk process will help to improve the computational complexity while maintaining the cluster quality.As a further measure for increasing the quality of the clusters,we plan to separate the meaningful data from statistic influences as proposed by Geyer-Schulz et al.

6.结论
我们已经展示了一个创新的方法,有效的建议,根据用户生成图书馆购买史与限制随机漫步.质量的建议以及计算算法的性能.
接下来的步骤,我们就会让你的接口公共为了获得用户反馈的建议并进行评估的基础上用户的判断.至于这个理论的算法,提出了调整参数会仔细观察.另一个重要的方面就是更新的建议按照图的变化引起的相似度的持续使用图书馆的网络接口.每个星期,图为更新和用法的历史时期.体积小,给予这些数据集的整体规模比较文件的数据,它是非常可取的更新程序,能够有效的将这些整合转化成簇,因此推荐名单.这个调查的人数的影响,从每走了节点以及更好的了解的渐近性态散步过程能够有效地提高了计算复杂度而维持集群质量.为进一步提高测量的质量,我们打算单独集群的有意义的数据统计影响从倡导的Geyer-Schulz等.