对中文文本聚类进行了研究,将所改进的DMK算法应用于实际的中文文本聚类中。将百度百科数据集中的不同类型词条内容经过文本处理(分词、去停用词、特征选取、降维)后分别使用原始K-means聚类算法和DMK(Density-based and Max-min-distan...对中文文本聚类进行了研究,将所改进的DMK算法应用于实际的中文文本聚类中。将百度百科数据集中的不同类型词条内容经过文本处理(分词、去停用词、特征选取、降维)后分别使用原始K-means聚类算法和DMK(Density-based and Max-min-distance K-means)算法进行聚类,并选择F-measure值及RI值(Rand Index)等指标对聚类结果进行分析。结果表明,针对实验中使用的百度百科中文数据集,DMK算法的F-measure值较原始算法平均提高0.342%,RI值较原始算法平均提高9.34%,验证了所设计的DMK算法对实际中文文本聚类的实质性优化。展开更多
研究了聚类分析技术在网络推荐系统中的应用。由于k均值(k-means)聚类算法易受到局部最优与噪声点等因素的影响,文章结合DBSCAN(Density-based Spatial Clustering of Application with Noise)算法和MMD(Max-Min Distance)初始聚类中心...研究了聚类分析技术在网络推荐系统中的应用。由于k均值(k-means)聚类算法易受到局部最优与噪声点等因素的影响,文章结合DBSCAN(Density-based Spatial Clustering of Application with Noise)算法和MMD(Max-Min Distance)初始聚类中心选取算法,对原始k-means算法进行了改进,提出了DMK(Density-based and Max-min-distance K-means)算法。该算法使用DBSCAN选取高密度点作为第一个聚类中心点的备选范围,接着选择相距最远的K-1个点作为其余的K-1个聚类中心,然后用得到的这组初始聚类中心进行k-means聚类。仿真与实验结果表明,该算法选择的初始聚类中心比较分散且代表性好,聚类的迭代次数减少,聚类结果的纯度提高。展开更多
A new bcu-MOF based upon a C2v-symmetrical[Cu_(4)Cl(COO)_(4)N_(8)]cluster,[Cu_(4)(Cl−)(L)_(4)·4H_(2)O]·(DMA^(+))·xDMF·yMeOH(SNNU-Bai66,SNNU-Bai=Shaanxi Normal University Bai's group),has been s...A new bcu-MOF based upon a C2v-symmetrical[Cu_(4)Cl(COO)_(4)N_(8)]cluster,[Cu_(4)(Cl−)(L)_(4)·4H_(2)O]·(DMA^(+))·xDMF·yMeOH(SNNU-Bai66,SNNU-Bai=Shaanxi Normal University Bai's group),has been sucessfully synthesized.By post-synthetic ion exchange experiments for the preparation of phases to be activated successfully,two new anionic derivatives,[Cu_(4)(Cl^(−))(L)_(4)·4H_(2)O]·[Li(H_(2)O)_(4)]^(+)·xDMK(Li^(+)-SNNU-Bai66,DMK=dimethyl ketone)and[Cu4(Cl−)(L)_(4)·_(4)H_(2)O]·(DMA^(+))·xDMK(DMA^(+)-SNNU-Bai66,DMA^(+)=Me_(2)NH_(2)),have been achieved and their selective CO_(2)or CH_(4)adsorptions over N_(2)gas have been systematically investigated in depth by both experimental and simulation studies.With the ionic pore surface and counterions cooperating with Lewis base N or O sites of the organic ligands in both Li^(+)-SNNU-Bai66 and DMA^(+)-SNNU-Bai66,the two MOFs exhibit highly selective CO_(2)and CH4 adsorption,especially for CH4 gas.Moreover,with the counterion being tuned from Li(H_(2)O)_(4)+to DMA^(+),selective CO_(2)adsorption properties are slightly optimized due to the higher polarizability of DMA^(+)and the preferred adsorption site changing from site I[the opposite side of Cl−toward Li(H_(2)O)_(4)^(+)]in Li^(+)-SNNU-Bai66 to site II(the position between the Hc atom of the L1 ligand and DMA^(+))in DMA^(+)-SNNU-Bai66,whereas those of CH_(4) of both MOFs are similar due to their having the same preferred adsorption site,namely,site I[the opposite side of Cl^(−)toward DMA^(+)or Li(H_(2)O)_(4)^(+)],in their frameworks.This work may further enrich the investigation of anionic MOFs in the field of selective gas adsorption with their inherent natures being more definitely displayed.展开更多
文摘对中文文本聚类进行了研究,将所改进的DMK算法应用于实际的中文文本聚类中。将百度百科数据集中的不同类型词条内容经过文本处理(分词、去停用词、特征选取、降维)后分别使用原始K-means聚类算法和DMK(Density-based and Max-min-distance K-means)算法进行聚类,并选择F-measure值及RI值(Rand Index)等指标对聚类结果进行分析。结果表明,针对实验中使用的百度百科中文数据集,DMK算法的F-measure值较原始算法平均提高0.342%,RI值较原始算法平均提高9.34%,验证了所设计的DMK算法对实际中文文本聚类的实质性优化。
文摘研究了聚类分析技术在网络推荐系统中的应用。由于k均值(k-means)聚类算法易受到局部最优与噪声点等因素的影响,文章结合DBSCAN(Density-based Spatial Clustering of Application with Noise)算法和MMD(Max-Min Distance)初始聚类中心选取算法,对原始k-means算法进行了改进,提出了DMK(Density-based and Max-min-distance K-means)算法。该算法使用DBSCAN选取高密度点作为第一个聚类中心点的备选范围,接着选择相距最远的K-1个点作为其余的K-1个聚类中心,然后用得到的这组初始聚类中心进行k-means聚类。仿真与实验结果表明,该算法选择的初始聚类中心比较分散且代表性好,聚类的迭代次数减少,聚类结果的纯度提高。
基金the National Natural Science Foundation of China(21771121)the Fundamental Research Funds for the Central Universities(No.GK201903040)for their support.
文摘A new bcu-MOF based upon a C2v-symmetrical[Cu_(4)Cl(COO)_(4)N_(8)]cluster,[Cu_(4)(Cl−)(L)_(4)·4H_(2)O]·(DMA^(+))·xDMF·yMeOH(SNNU-Bai66,SNNU-Bai=Shaanxi Normal University Bai's group),has been sucessfully synthesized.By post-synthetic ion exchange experiments for the preparation of phases to be activated successfully,two new anionic derivatives,[Cu_(4)(Cl^(−))(L)_(4)·4H_(2)O]·[Li(H_(2)O)_(4)]^(+)·xDMK(Li^(+)-SNNU-Bai66,DMK=dimethyl ketone)and[Cu4(Cl−)(L)_(4)·_(4)H_(2)O]·(DMA^(+))·xDMK(DMA^(+)-SNNU-Bai66,DMA^(+)=Me_(2)NH_(2)),have been achieved and their selective CO_(2)or CH_(4)adsorptions over N_(2)gas have been systematically investigated in depth by both experimental and simulation studies.With the ionic pore surface and counterions cooperating with Lewis base N or O sites of the organic ligands in both Li^(+)-SNNU-Bai66 and DMA^(+)-SNNU-Bai66,the two MOFs exhibit highly selective CO_(2)and CH4 adsorption,especially for CH4 gas.Moreover,with the counterion being tuned from Li(H_(2)O)_(4)+to DMA^(+),selective CO_(2)adsorption properties are slightly optimized due to the higher polarizability of DMA^(+)and the preferred adsorption site changing from site I[the opposite side of Cl−toward Li(H_(2)O)_(4)^(+)]in Li^(+)-SNNU-Bai66 to site II(the position between the Hc atom of the L1 ligand and DMA^(+))in DMA^(+)-SNNU-Bai66,whereas those of CH_(4) of both MOFs are similar due to their having the same preferred adsorption site,namely,site I[the opposite side of Cl^(−)toward DMA^(+)or Li(H_(2)O)_(4)^(+)],in their frameworks.This work may further enrich the investigation of anionic MOFs in the field of selective gas adsorption with their inherent natures being more definitely displayed.
