GLSLIM模型(Global and Local SLIM)是基于SLIM模型(Sparse Linear Methods)并优于SLIM模型的推荐算法。它将GLOBAL模型和LOCAL模型结合,GLOBAL模型用来捕获物品在所有用户之间的相似性,LOCAL模型用来获取物品在某个用户子集中的相似性...GLSLIM模型(Global and Local SLIM)是基于SLIM模型(Sparse Linear Methods)并优于SLIM模型的推荐算法。它将GLOBAL模型和LOCAL模型结合,GLOBAL模型用来捕获物品在所有用户之间的相似性,LOCAL模型用来获取物品在某个用户子集中的相似性,通过两种模型的结合可以进一步优化用户的个性化推荐。但该算法存在天然缺陷,就是被用户评价或购买过的物品之间的相似度才可以被学习到,没有被购买过的物品之间的相似度为0。这将导致用户购买过的相似物品才有机会被推荐,相似度为0的物品无法推荐给用户。为了改善这种情况,利用一种PLSA模型解决这个问题,基于两种模型的组合进行协同过滤推荐。实验结果表明,虽然推荐结果的准确性略微降低,但是能挖掘用户的潜在兴趣。展开更多
Depend on the good performance, the CsI crystal developed by the Crystal Detector Group has a good reputation in the international market. So far, In addition to meeting the needs of the external target experimental t...Depend on the good performance, the CsI crystal developed by the Crystal Detector Group has a good reputation in the international market. So far, In addition to meeting the needs of the external target experimental terminal in the CSR, pieces of CsI(Tl) crystals have been sold to some international and domestic research institutes, colleges and universities. They have been used for kinds of experiments in some well-known labs in German, America and Japan, and so on.展开更多
文摘GLSLIM模型(Global and Local SLIM)是基于SLIM模型(Sparse Linear Methods)并优于SLIM模型的推荐算法。它将GLOBAL模型和LOCAL模型结合,GLOBAL模型用来捕获物品在所有用户之间的相似性,LOCAL模型用来获取物品在某个用户子集中的相似性,通过两种模型的结合可以进一步优化用户的个性化推荐。但该算法存在天然缺陷,就是被用户评价或购买过的物品之间的相似度才可以被学习到,没有被购买过的物品之间的相似度为0。这将导致用户购买过的相似物品才有机会被推荐,相似度为0的物品无法推荐给用户。为了改善这种情况,利用一种PLSA模型解决这个问题,基于两种模型的组合进行协同过滤推荐。实验结果表明,虽然推荐结果的准确性略微降低,但是能挖掘用户的潜在兴趣。
文摘Depend on the good performance, the CsI crystal developed by the Crystal Detector Group has a good reputation in the international market. So far, In addition to meeting the needs of the external target experimental terminal in the CSR, pieces of CsI(Tl) crystals have been sold to some international and domestic research institutes, colleges and universities. They have been used for kinds of experiments in some well-known labs in German, America and Japan, and so on.
