文献搜索引擎中特征项及权重的应用

针对目前用户在使用搜索引擎过程中,检索结果冗余、效率低下等问题,本文在对文献垂直搜索系统中Lucene.Net的索引算法研究基础上,结合用户实际专业检索需求,改变其算法中的激励因子,实验结果证明该方法确能提高搜索结果的相关度.     

增压风机和引风机协调控制策略的应用

脱硫系统增压风机一般用来控制增压风机入口压力,由于风烟系统的管路特性,炉膛压力变化和增压风机入口压力变化存在一定的时间差,当烟气流量大幅变动时,增压风机控制往往容易和引风机的炉膛压力控制产生谐振,导致被控量波动幅度过大,波动时间过长,甚至发生发散。为了克服这种现象,在传统逻辑中针对风机RB增加了相应的前馈,并减弱了增压风机的调节作用,从而降低了正常工况下增压风机的控制品质。针对上述问题,本文通过理论分析,在仿真的基础上提出了一种风机协调控制的控制策略。并以某300MW机组的工程实例,说明了该控制策略的实施过程和实施效果。     

基于二次TF* IDF的互信息文本特征选择算法研究

在分析传统互信息法缺陷的基础上,提出一种基于二次TF*IDF的互信息特征选择算法,对仅在一个类别中出现的特征词的重要程度给予再次的衡量,解决了互信息值相等而无法进行有效特征选择的问题。利用贝叶斯分类器对该方法进行验证的结果表明该算法在文本分类效率和正确率上比原有方法有一定的提高。     

Statistical evaluation on storm sewer design criteria under climate change in Seoul,South Korea

This study conducted statistical analysis under both stationary and non-stationary climate using rainfall data measured in Seoul, South Korea. The 50 year annual maximum rainfall data were analysed with generalised extreme value distribution and Gumbel distribution. Possible rainfall changes were predicted with nonstationary forms of the two distributions, considering time variable location parameters. The current storm sewer design criteria were assessed by reflecting climate change implications and expected lifetime performance. Expected return periods were reduced to 4.67, 8.66, 19.16 and 23.53 years for the current 10 min - 5, 10, 30 and 50 years. The newly suggested design criteria of 5, 10, 30 and 50 yr return period storm events were 5.3, 12, 48 and 107 years for the 10 min duration for a 50 year lifetime expectancy.     

高压变频调速在火电厂锅炉引风系统中的应用

本文介绍了利用TMEIC的MV系列变频器对火电厂锅炉引风系统进行改造，对电力电子技术、异步电机变频的控制方式和MV系列变频器的工作原理及特点等进行了介绍，对改造后的系统特点、经济效益进行了分析，指出改造后降低了煤耗、节省了电能，并可在火电厂其它平方转矩负载上的推广应用。     

PCB喷淋系统应用浅谈

文章针对PCB厂水平线如何计算、选择合适的喷嘴进行了初步的探讨,并结合实际应用展示了选择合适的喷淋系统,以有利于生产品质的提升。     

高压变频器在600MW机组引风机的应用

就大容量高压变频器在600MW机组引风机的应用进行综合分析。从高压变频器的结构组成、性能指标分析及现场安装调试进行详细阐述,并对改造后的节能进行客观分析评价。结合运行实际,分析了高压变频器运行后对风机本体的影响及存在问题,有针对性地提出防范措施。     

New doctors ranking system based on VIKOR method

Nowadays, we can use different websites that help us make decisions about various aspects of our lives. However, privacy protection prevents websites from providing personalised guidelines to users. We propose a novel doctor‐ranking system (DRS) based on multi‐criteria group decision‐making (MCGDM) method to address the problems of privacy protection. The following aspects differentiate our proposed DRS model from previous works: (a) textual information reviews are used to identify user preferences and complementary criteria, (b) criteria weights are determined by term frequency inverse document frequency (TF‐IDF) instead of Delphi method or expert opinion, (c) intuitionistic fuzzy sets (IFSs) are used to replace sentiment analysis to express subjective user criteria, and (d) VIsekriterijumsko KOmpromisno Rangiranjie (VIKOR) method for MCGDM with IFSs is used to solve the doctor‐ranking problem. We apply our proposed model to datasets from Haodf.com to compare the performance of our method with that of sentiment analysis and technique for order performance by similarity to ideal solution (TOPSIS) methods. The experimental results show that our method provides accurate ranking and increases the reliability of DRS.     

基于双阈值Apriori算法和非频繁项集的关联规则挖掘方法

针对从本文数据集中的正负关联规则挖掘问题,提出一种基于双阈值Apriori算法和非频繁项集的挖掘方法。首先,对通过逆文档频率(IDF)对语料库中的项(项集)进行加权,筛选出前N%的项集。然后,通过提出的双支持度阈值Apriori算法来提取频繁项集和非频繁项集,以此降低非频繁项集的数量。最后,通过置信度和升降度阈值的判断,分别从频繁项集和非频繁项集中挖掘正负关联规则。其中,创新性的利用了非频繁项集来挖掘正负关联规则。在一个医学文本数据集上的实验结果表明,提出的方法能够有效挖掘出正负关联规则,且能够大大降低项集和规则数量。