Page 155 - JOURNAL OF LIBRARY SCIENCE IN CHINA 2018 Vol. 42

154 Journal of Library Science in China, Vol. 8, 2016

Table 4. Method disadvantage description rules and examples
Rule Example
但面对高维稀疏的数据，近邻传播算法往往不能得到很好的聚类结果，而且该方法不能产
…不能|无法|… 生指定类数的聚类
*** cannot When the data are both high-dimensional and sparse, nearest propaganda algorithms cannot achieve
a good clustering, and the method cannot generate a certain number of clusters either
而该方法存在两个明显的缺点，其一是在训练统计翻译模型t(q i |w)时，需要大量的训练数
据，而这在实际应用中很难获得
The method has two apparent disadvantages; one is that it requires a large amount of training data,
which is difficult to get in practice when trying to train the statistical translation model t(q i |w)
…，[但|可是]存在| 这种方法的优点是容易实现，但它有两方面的缺点，一是某些常用的单个标签缺少语义的
尚存|具有…的缺点| 明确性，不能体现用户的个性化偏好，如标签“Web”等；二是用户在对具体对象进行标签
不足 时，同时使用了多个标签，它们之间存在着一定的语义联系，向量型的离散表示不能刻画
*** has *** disad- 标签之间的这种语义联系
vantages The advantage of this method is that it is easy to implement the method.However it also has two
disadvantages. One disadvantage is that common labels, like “web”, lack the clarity on semantics
and cannot represent the user’s preference. Another disadvantage is that if a user annotates a
certain object using multiple labels, it implies that certain relationships between the labels may
exist. A discrete representation like a vector cannot characterize this sort of semantic relation
…极易 + 造成 | 带 现有的混合多准则决策问题的解决方法极易造成准则信息的丢失，同时，计算和转换的过
来… 程过于复杂
*** can easily Current hybrid MCDM problems can easily result in the loss of information guidelines while the
result in *** computation and transform are complex

Table 5. Method advantage description rules and examples
Rule Example
共现分析是基于定量化的数据分析，以其方法的简明性和分析结果的可靠性成为支撑信息
…以…成为|为… 内容分析研究过程的重要手段和工具
的… Co-occurrence analysis is based on quantitative data analysis, andits simplicity and reliability of
*** makes ***
analytical results makes it an important means and tool for supporting content analysis process
对于…，…是一
种…的方法 对于处理大规模的数据集，近邻传播算法是一种快速、有效的聚类方法
For ***, *** is a For handling large data sets, affinity propagation algorithm is a fast and effective clustering method
*** method
采用信息熵的参数确定方法综合了计算法和实验法的优点，操作简单、方便，更能反映出
…方法+综合了|兼 参数与样本联系的本质特征，客观的刻画了样本的重要性程度
顾了…的优点，… Determining the parameters using information entropy method combines the advantages of
*** combines the computation and the experimental method. The method is simple, convenient and better reflects the
advantages of *** essential characteristics of the parameters associated with the sample and objectively portrays the
importance of the sample
由于本文方法不需要先验知识，通过对数据集的全局进化优化，输出规模更小，结构简单便于
理解的规则集，通过否定选择分类器进行规则匹配，克服了其仅适合样本稀疏或小样本空间的
…方法+克服了|避 问题，在分类效果上优于决策树Assistant R方法，同时在分类效率上高于K-NN、贝叶斯方法
免了…的缺点，… Since this method does not require prior knowledge, through the global optimization of the data
The method over- set, we get smaller output size, set of rules, that hasa simple structure and is easy to understand.
comes the advantage Negative selection classifier match rule overcomes the disadvantages of adapting to only the sparse
of ***
samples or small sample problem space. The method achieves better results compared with decision
tree Assistant R method and is more effective than the K-NN, Bayesian approach
与…相比，…方法 与传统聚类相比较，三维空间聚类法最大好处是直观、灵活，能够很好地揭示词义间内在
[在…方面]+具有 联系
更…的… Compared with traditional clustering methods, the merits of 3-dimension clustering is that it is
Compared with ***, intuitive, flexible and can represent the intrinsic relationship between semantics
***ismore