国外网络分类法的发展及对我国网络分类法理论与实践的启示

论文在分析网络分类法与传统文献分类法差异的基础上，阐述了国外网络分类法的特点、问题、发展趋势以及对我国网络分类法理论与实践的几点启示。     

基于粗糙集加权的文本分类方法研究

文本自动分类是当前智能信息处理中一类重要的研究课题。本文分析了基于统计理论的文本分类的基本特点,提出采用可变精度粗糙集模型中的分类质量构造新的特征词权重计算公式。这种新的加权方法,相对于广泛使用的逆文本频率加权方法,大大改进了文本样本在整个空间中的分布,使得类内距离减少,类间距离增大,在理论上将提高样本的可分性。最后利用支持向量机和K近邻两种分类器,验证了这种新的加权方法对分类效果确实有所提高。     

简评三段式分类索刊号——与施振宏、王觊商稚兼谈刊次号设计

针对三段式分类索刊号的编号方法及在实际应用中存在的不足提出商讨意见，并对刊次号的设计提出看法。     

Analysis of Statistical Question Classification for Fact-Based Questions

Question classification systems play an important role in question answering systems and can be used in a wide range of other domains. The goal of question classification is to accurately assign labels to questions based on expected answer type. Most approaches in the past have relied on matching questions against hand-crafted rules. However, rules require laborious effort to create and often suffer from being too specific. Statistical question classification methods overcome these issues by employing machine learning techniques. We empirically show that a statistical approach is robust and achieves good performance on three diverse data sets with little or no hand tuning. Furthermore, we examine the role different syntactic and semantic features have on performance. We find that semantic features tend to increase performance more than purely syntactic features. Finally, we analyze common causes of misclassification error and provide insight into ways they may be overcome.     

基于用户自然标注的TF-IDF辅助标引算法及实证研究

[目的/意义] 从用户角度出发,研究基于用户自然标注的TF-IDF辅助标引算法。[方法/过程] 首先以核心期刊论文中作者标注的关键词和分类号为源数据,通过对关键词词频进行统计,使用TF-IDF算法构建用户标注词表、形成标引知识库,然后通过IK Analyzer分词软件对待标引的科技项目数据进行切词和停用词处理,进而使用TF-IDF算法和位置加权算法提取科技项目数据的特征词,最终实现对科技项目数据进行关键词和分类的同步标引。[结果/结论] 实验结果表明,机标关键词与人标关键词的相似比在60%以上的科技项目数据占总数的68.1%,机标分类号与人标分类号前三位一致的占总数的83.9%,结果表明基于用户自然标注数据并采用TF-IDF算法在关键词和分类标引方面是可行的。     

Granularity of algorithmically constructed publication-level classifications of research publications: Identification of topics

The purpose of this study is to find a theoretically grounded, practically applicable and useful granularity level of an algorithmically constructed publication-level classification of research publications (ACPLC). The level addressed is the level of research topics. The methodology we propose uses synthesis papers and their reference articles to construct a baseline classification. A dataset of about 31 million publications, and their mutual citations relations, is used to obtain several ACPLCs of different granularity. Each ACPLC is compared to the baseline classification and the best performing ACPLC is identified. The results of two case studies show that the topics of the cases are closely associated with different classes of the identified ACPLC, and that these classes tend to treat only one topic. Further, the class size variation is moderate, and only a small proportion of the publications belong to very small classes. For these reasons, we conclude that the proposed methodology is suitable to determine the topic granularity level of an ACPLC and that the ACPLC identified by this methodology is useful for bibliometric analyses.     

Four pitfalls in normalizing citation indicators: An investigation of ESI’s selection of highly cited papers

InCites Essential Science Indicators is becoming increasingly used to identify top-performing research and evaluate the impact of institutes. Unfortunately, our study shows that ESI indicators, as well as other normalized citation indicators, have the following flaws. First, the publication month and the online-to-print delay affect a paper’s probability of becoming a Highly Cited Paper (HCP). Papers published in the earlier months of the year are more likely to accumulate enough citation counts to rank at the top 1% compared with those published in later months of the year. Papers with longer online-to-print delays have an apparent advantage for being selected as HCPs. Research field normalizations lead to the third pitfall. Different research fields have different citation thresholds for HCPs, making research field classification important for a journal. In addition, the uniform thresholds for both articles and reviews in ESI affect the reliability of HCP selection because, on average, reviews tend to have higher citation rates than articles. ESI’s selection of HCPs provides an intuitive feel for the problems of normalized citation impact indicators, such as those provided in InCites and SciVal.     

数字调制类型的识别及实现

通信信号的调制识别是目前研究的热点问题.本文给出一种新的数字调制信号识别方法.高速数字专用芯片ADSP2101是此方法具体实现的重要器件,这一方法在实际通信中有着广泛的应用.     

Writer identification approach based on bag of words with OBI features

Handwriter identification aims to simplify the task of forensic experts by providing them with semi-automated tools in order to enable them to narrow down the search to determine the final identification of an unknown handwritten sample. An identification algorithm aims to produce a list of predicted writers of the unknown handwritten sample ranked in terms of confidence measure metrics for use by the forensic expert will make the final decision.Most existing handwriter identification systems use either statistical or model-based approaches. To further improve the performances this paper proposes to deploy a combination of both approaches using Oriented Basic Image features and the concept of graphemes codebook. To reduce the resulting high dimensionality of the feature vector a Kernel Principal Component Analysis has been used. To gauge the effectiveness of the proposed method a performance analysis, using IAM dataset for English handwriting and ICFHR 2012 dataset for Arabic handwriting, has been carried out. The results obtained achieved an accuracy of 96% thus demonstrating its superiority when compared against similar techniques.     

An NCME Instructional Module on Data Mining Methods for Classification and Regression

Data mining methods for classification and regression are becoming increasingly popular in various scientific fields. However, these methods have not been explored much in educational measurement. This module first provides a review, which should be accessible to a wide audience in education measurement, of some of these methods. The module then demonstrates using three real‐data examples that these methods may lead to an improvement over traditionally used methods such as linear and logistic regression in educational measurement.