基于隐特征提取和CRF模型的中文词义消歧

传统的中文词义消歧方法是通过观察文本的上下文信息、词性等显性特征建立消歧模型,本文通过对歧义产生原因进行深入的分析,发现词语之间隐含的语法结构、语义信息等也会导致歧义的产生,可以考虑将这些信息加入消歧模型进行消歧。由于《知网》知识库中对词语之间的搭配信息进行了总结,本文借助《知网》提取训练语料库所获取的词语搭配信息的隐性语义特征,结合显性的上下文特征,采用条件随机场的方法进行词义消歧。最后,通过实验进行词义消歧和效果验证,结果表明：本文采用的方法与传统的条件随机场消歧相比,词义消歧的准确率得到了提高。     

基于NER的石油非结构化信息抽取研究

随着"智能油田"的建设加快,构建基于海量石油数据的智能分析系统意义重大。然而,由于石油生产过程中产生的文本数据往往无结构且类型多样,从中抽取关键信息进行分析成为一个研究热点,而信息抽取又需要高质量的语义实体做支撑。根据这一特定问题,提出基于命名实体识别（Named Entity Recognition,NER）技术针对石油非结构化文本进行信息抽取,构建双向长短时记忆（Bidirectional Long Short-Term Memory,Bi LSTM）网络模型提取语料特征,并结合条件随机场（Conditional Random Field,CRF）做分类器,构建了基于Bi LSTM+CRF的高精度NER模型,针对石油工业领域的非结构化文本进行命名实体抽取。通过在修井作业文本数据集上进行对比实验表明,本方法具有较高的精确率和召回率。     

高密度间苯二酚-甲醛碳气凝胶ICF靶的制备与吸附性能研究

 以间苯二酚和甲醛为前驱体,通过改进传统制备技术解决了高密度间苯二酚-甲醛(CRF)碳气凝胶制备过程中的龟裂问题,制备出了符合ICF实验需要的高密度CRF碳气凝胶材料。分别对CRF碳气凝胶的元素组成和物相组成进行了鉴定,采用自动吸附仪考察了CRF碳气凝胶对N₂和H₂的吸附性能。结果表明：该碳气凝胶是一种由C元素组成的类似石墨结构的非晶固态材料,结构均匀性好,具有良好的机械加工性能,比表面积达676 m²·g^-1,平均孔径为7.16 nm;氢吸附质量分数达2.28%,相应体积密度为17.83 kg·m^-3。     

Recognition of Chinese Organization Name Using Co-training

Chinese organization name recognition is hard and important in natural language processing. To reduce tagged corpus and use untagged corpus, we presented combing Co-trainins with support vector machines (SVM) and conditional random fields (CRF) to improve recognition results. Based on principles of uncorrelated and compatible, we constructed different classifiers from different views within SVM or CRF alone and combination of these two models. And we modified a heuristic untagged samples selection algorithm to reduce time complexity. Experimental results show that under the same tagged data, Co-training has 10% F-measure higher than using SVM or CRF alone; under the same F-measure, Co-training saves at most 70% of tagged data to achieve the same performance.     

Advancement in optimization tactic achieved by newly developed chromatographic response function: application to LC separation of raloxifene and its impurities

In this paper a new chromatographic response function (CRF) is designed and proposed for utilization in the optimization strategies. The function capability to represent the overall quality of a experimentally obtained chromatograms was compared to the other two objective functions and proved to give more accurate and reliable results. The new CRF has improved concept of separation and time term estimation. It reflects all important defects of the chromatogram such as the appearance of asymmetrical or overlapping peaks and prolonged elution time and allows the appropriate weighting of each of them. The LC separation of raloxifene and its four impurities was evaluated through the central composite design experimental plan choosing the new CRF to be the only output of the system. The function demonstrated the ability to judge the impact of the complex interactions of the selected chromatographic parameters (acetonitrile content in the mobile phase, sodium dodecyl sulfate concentration in the water phase, pH of the mobile phase and column temperature) on the mixture behavior and led to the determination of the optimal separation conditions. The newly developed CRF proved to have the advanced performances and it presents the important step forward in the optimization of the chromatographic separation.     

现浇钢筋混凝土暗梁空心板在某建筑中的应用

介绍了现浇暗梁空心楼盖技术,结合帝景花园商场工程对结构梁板的设计计算进行了说明。     

基于主题标签和CRF的中文微博命名实体识别

近年来,网络媒体微博的迅速发展,为命名实体的识别研究提供了一种全新的载体.针对中文微博文本短、表达不清、网络化严重等特点,论文提出了一种规则与统计相结合的中文微博命名实体识别方法.该方法首先利用中文微博的主题标签对处理后的数据进行筛选,然后再选取合适的特征模板,并利用条件随机场模型(Conditional random fields, CRF)进行实体识别.为了满足实验要求,该文将传统网页爬虫方法与API接口采集方法相结合进行微博数据采集.实验结果表明,该方法能够有效提高中文微博命名实体的识别效果.     

基于“治未病”理念下的四维一体综合干预技术防治CRF临床研究

目的探讨自拟中医药口服方,灌肠方,沐浴方及穴位按摩联合运用的四位一体,全方位,多靶点综合干预技术,观察治疗CRF患者的临床疗效。方法符合纳入CRF标准的患者随机分为治疗组（实施中医药四位一体干预技术组）和对照组（口服西药包醛氧淀粉组）：观察二组患者客观症状的改善,肌酐（Scr）、尿素氮（Bun）,内生肌酐清除率（Ccr）等生化指标变化,评定疗效。结果经3个疗程90d的治疗比较,治疗组和对照组CRF患者的Scr,Bun,Ccr变化进行比较有极其显著差异P〈0．01）。结论四位一体综合干预技术防治CRF疗效显著,值得临床推广应用。     

Evaluation of Seven Chromatographic Response Functions on Simulated and Experimentally Obtained Chromatograms in Hydrophilic Interaction Liquid Chromatography System

This paper investigates the ability of seven chromatographic response functions to measure the quality of chromatograms obtained in hydrophilic interaction liquid chromatography (HILIC). First, the functions were tested on a set of simulated chromatograms and differences in their mathematical design were discussed. Second, the functions were evaluated on the experimentally obtained chromatograms in HILIC analysis of model mixture consisted of beta agonists and antagonists. The ranking of chromatograms obtained by different functions was significantly different, implying that the accuracy of the optimization procedure is strongly dependent on the function that was selected as an output. Investigation of potential drawbacks of each function was conducted and general recommendations concerning the use of chromatographic response functions in optimization strategies are proposed.     

基于深度主动学习的信息安全领域命名实体识别研究

针对通用领域模型不能很好地解决信息安全领域的命名实体识别问题,提出一种基于字符特性,双向长短时记忆网络(Bi-LSTM)与条件随机场(CRF)相结合的信息安全领域命名实体识别方法.该方法不依赖于人工选取特征,通过神经网络模型对序列进行标注,再利用CRF对序列标签的相关性进行约束,提高序列标注的准确性.而且,针对信息安全领域标注数据样本不足的问题,采用主动学习方法,使用少量标注样本达到较好的序列标注效果.