Improved Term Weighting Technique for Automatic Web Page Classification

Automatic web page classification has become inevitable for web directories due to the multitude of web pages in the World Wide Web. In this paper an improved Term Weighting technique is proposed for automatic and effective classification of web pages. The web documents are represented as set of features. The proposed method selects and extracts the most prominent features reducing the high dimensionality problem of classifier. The proper selection of features among the large set improves the performance of the classifier. The proposed algorithm is implemented and tested on a benchmarked dataset. The results show the better performance than most of the existing term weighting techniques.

Supervised Contrastive Learning with Term Weighting for Improving Chinese Text Classification

With the rapid growth of information retrieval technology,Chinese text classification,which is the basis of information content security,has become a widely discussed topic.In view of the huge difference compared with English,Chinese text task is more complex in semantic information representations.However,most existing Chinese text classification approaches typically regard feature representation and feature selection as the key points,but fail to take into account the learning strategy that adapts to the task.Besides,these approaches compress the Chinese word into a representation vector,without considering the distribution of the term among the categories of interest.In order to improve the effect of Chinese text classification,a unified method,called Supervised Contrastive Learning with Term Weighting(SCL-TW),is proposed in this paper.Supervised contrastive learning makes full use of a large amount of unlabeled data to improve model stability.In SCL-TW,we calculate the score of term weighting to optimize the process of data augmentation of Chinese text.Subsequently,the transformed features are fed into a temporal convolution network to conduct feature representation.Experimental verifications are conducted on two Chinese benchmark datasets.The results demonstrate that SCL-TW outperforms other advanced Chinese text classification approaches by an amazing margin.

Fetal Thigh Circumference versus Fetal Abdominal Subcutaneous Tissue Thickness in Prediction of Fetal Weight in Term Pregnant Women

Background: Fetal weight estimation by ultrasound is an important factor in obstetrics;it is directly related to the gestational age which helps to plan the mode of delivery and labor management. Objective: to compare between fetal thigh circumference (TC) and abdominal subcutaneous tissue thickness (SCT) in estimating birthweight in term pregnant women. Patients and Methods: This prospective cohort study was conducted at outpatient clinic or emergency room, Obstetrics and Gynecology Department, Faculty of Medicine, Ain Shams University Maternity Hospitals from March 2022 until May 2023. During this study, 100 term pregnant females with gestational age 37 - 40 weeks attended El Demerdash Maternity Hospital and scheduled for delivery either at outpatient clinic or emergency room were enrolled, after consenting each of them. Basic fetal biometry was performed by an expert and professional medical personnel to ensure the accuracy of examination results. Fetal abdominal subcutaneous thickness and fetal thigh circumference were measured for assessment of gestational age and correlated with actual fetal body weight. In the current study, three formulas;Hadlock, Vintzileos’ and SCT formula were correlated with actual fetal body weight after birth. Results: The present study revealed that TC formula is closer to the actual birth weight, followed by Hadlock formula, while the SCT formula is the furthest from it. Conclusion: To increase the accuracy of birth estimations, regular ultrasound examinations should include fetal thigh circumference measurement.

Relationships of Surfactant Head Group Weight Fraction and Some Polarity Terms by Gas Chromatography

Polarity, partition coefficient （K）, methanol carbon number of surfactant ((CMeOH)S), and methanol carbon number of surfactant head group ((CMeOH)HG) are measured on six alkanolamides and five polyoxyethylenated long chain amines as stationary phases. From the measured methanol carbon numbers, polarity indices, （IP）S and （IP）HG, are calculated. The determined polarity terms are plotted against the head group weight fraction ( fHG ) of the investigated surfactants and several equations have been developed. The study reveals that the molecular structural gap between alkanolamides and polyoxyethylenated long chain amines diminished when HLB numbers of these surfactant classes are plotted against fHG values. Consequently, a general equation relating HLB and fHG is obtained.

Prevalence of term low birth weight andits determinants in Shanghai, China

Objective: To investigate the prevalence of term low birth weight (TLBW) and its risk factors. Methods: A follow-up study with 7, 872 couples was conducted from 1987 to 199o beginning from the time they got marriage licenses in two districts defined in Shanghai. They were interviewed in the third month and again in the fifteenth month and in the fifth to sixth year afterwards individually at home. The total follow up rate reached 98%. Couple’s background characteristics as well as the information on their general health. reproductivc history and contraceptive use etc.. were collected dynamically. All of the single live births with term delivery were Included for data analysis in this paper. Adjusted odd ratios and population attributable risk (PAR%) were computed. Results: The prevalence of TLBW in Shanghai single term live births was 2. 0% (134,/6.573), represents 54. 7% (134/245) of the total low birth weights in our sam pie. Significant social and behaviour risk factors relating with TI-BW were wife’s dissat- isfaction with marriage; low education level of husband; co-residence with parents during pregnancy; heavy housework done by the wife while being pregnant. Significant biomedical risky factors were menarche age greater than 16 years old; maternal age at delivery greater than 29 years old; maternal body mass index less than 19. 8; wife suf- fered from serious disease prior to conceiving; having pregnancy complication; gestational weight gain less than 20 % of pre-pregnancy weight; having abortion, stillbirth and fetal death history. Conclusion: TL.BW constituted over half of all low birth weights in Shanghai. Special attention should be paid to the determinants mentioned above in TLBW intervention program. Improving couples’ economic and living condition and husband ’s education at tainment, and caloric supplementation with women while being pregnant would all be particularly effective in reducing the occurrence of TLBW in Shnaghai.

Changes in weight and metabolic parameters following long-term iloperidone use: A meta-analysis of data from 9 Phase II and III trials of iloperidone

Objective: To determine the impact of iloperidone treatment on metabolic parameters. Methods: Data were analyzed from 3210 patients who received iloperidone and 587 patients who received placebo for up to 2 years during 9 Phase II and Phase III studies (integrated safety database). Patients were pooled by most frequent iloperidone dose into 3 groups: 4 - 8 mg/d, 10 - 16 mg/d, and 20 - 24 mg/d. Laboratory data from 8 studies were random (fasting and nonfasting) while patients from 1 study (n = 447) were all sampled at fasting. Results: Mean (standard deviation [SD]) weight gain from baseline to endpoint was 2.1 (6.8) kg for all patients on iloperidone. Fasting mean (SD) total cholesterol changes from baseline to end of study were 8.2 (31.6) for iloperidone and -2.2 (35.2) mg/dL for placebo. Fasting mean (SD) triglyceride changes during this period were iloperidone: -0.83 (82.3);placebo: 16.5 (113.1) mg/dL. Fasting mean (SD) glucose changes were iloperidone: 6.6 (24.0);placebo: -0.05 (17.0) mg/dL from baseline to end of study. Glycosylated hemoglobin levels were unchanged following iloperidone treatment. Conclusion: These data suggest a metabolic profile for iloperidone of modest weight gain accompanied by small changes in lipids and glucose that are unlikely to be of clinical concern.

Multiple Solutions to the Problem of Kirchhoff Type Involving the Critical Caffareli-Kohn-Niremberg Exponent, Concave Term and Sign-Changing Weights

In this paper, we establish the existence of at least four distinct solutions to an Kirchhoff type problems involving the critical Caffareli-Kohn-Niremberg exponent, concave term and sign-changing weights, by using the Nehari manifold and mountain pass theorem.

基于VMD-改进最优加权法的短期负荷变权组合预测策略

为提升短期电力负荷预测精度,提出了一种变权组合预测策略。首先,为了降低负荷数据的不平稳度,使用变分模态分解(variational mode decomposition,VMD)将负荷数据分解成了高频、低频、残差3种特征模态分量。其次,充分计及负荷数据的时序特点,参考指数加权法原理设计自适应误差重要性量化函数,并结合组合模型在时间窗口内的历史负荷数据的均方预测误差设计改进最优加权法的目标函数和约束条件,以完成子模型的准确变权。最后,针对波动较强的高频分量选定极端梯度提升(XGBoost)和卷积神经网络-长短期记忆(CNN-LSTM)模型并使用改进最优加权法进行组合预测、低频分量使用多元线性回归(MLR)模型预测、残差分量使用LSTM模型预测,叠加各模态分量的预测结果,实现了短期负荷数据的准确预测。实验结果表明,使用策略组合模型的平均绝对百分比误差为4.18%。与使用传统组合策略的组合模型相比,平均绝对百分比预测误差平均降低了0.87%。

面向语法加权图文本的方面情感三元组抽取

方面情感三元组抽取包括方面抽取、意见抽取和方面情感分类3项任务,以管道方式解决该任务的研究方法无法利用元素之间的交互信息,同时也会造成错误传播和冗余训练。基于此,提出一种基于门控注意力和加权图文本的方面情感三元组抽取方法。采用双向长短时记忆网络学习句子的序列特征表示;利用门控注意力单元学习单词之间的线性联系;利用语法距离加权图卷积网络增强三元组元素之间的交互;利用网格标记推理策略预测三元组。在4个公开数据集上进行实验,结果表明:所提方法可以有效增强三元组元素之间的交互,提高三元组抽取的准确率;同时,所提方法的F1值分别为57.94%、70.54%、61.95%和67.66%,与基准模型相比均有所提高。

基于分类的term重要性识别方法

在传统的搜索引擎和信息检索中,用户Query中的term-weight通常是以一种上下文无关的方式得到的。现有的大多数信息检索技术都使用词袋方法,例如布尔模型、向量空间模型和概率模型等,这些方法均没有考虑Query中term之间的相关性。为了能够充分利用Query中的信息来提高term-weight的准确度,提出了一种有监督的机器学习方法来学习用户Query中的term-weight。该方法基于分类的方法,并引入了句法分析作为分类的一项重要的特征来训练模型。考虑用户Query中term之间的关系后,既避免了由Query到单个term的信息丢失,又增加了短文本的特征,同时使分类器实现软输出,能够给term的重要程度一个更为准确的量化值。

基于SWLSTM-Stacking集成学习的源荷区间预测方法

针对光伏出力和电动汽车充电特性的随机特性对电力系统的冲击不断增强,准确及时的源荷预测是实现增强电力系统适应性和稳定性的重要课题。因此,提出一种基于共享权重长短期记忆网络(shared weight long short-term networks,SWLSTM)与Stacking集成模型相结合的源荷区间预测方法。首先,光伏出力存在时序性特征,采用局部线性嵌入改进k-means算法聚类提取特征日,在实现数据降维同时,减少了网络训练难度;其次,在Stacking集成模型的框架下,将SWLSTM作为元学习器,并通过Q统计量筛选合适的基学习器模型,从而实现多模型融合的多异学习器Stacking集成学习的源荷预测;紧接着,为了得到预测的不确定信息,引入置信度区间预测;最后,采用实测数据对本文所提方法进行验证。结果表明改进k-means算法能够降低其求解难度,加快求解速度,可以快速获取聚类特征;所引入集成学习模型和置信度区间,有效表征源荷预测的不确定性,提升区间预测模型的泛化能力。

CEEMDAN-WPE-CLSA超短期风电功率预测方法研究

提出了一种结合自适应噪声完全集合经验模态分解、加权排列熵、卷积神经网络、长短期记忆网络和自注意力机制的超短期风电功率预测方法。首先,利用自适应噪声完全集合经验模态分解将原始风电功率时间序列自适应分解为一系列的模态分量,降低原始序列的非线性和波动性;其次,根据加权排列熵计算各模态分量间的相似性并对相似的分量进行重组,以修正自适应噪声完全集合经验模态分解的过度分解问题,使得修正后的模态分量更具规律性;最后,将重组后的分量输入卷积长短期记忆网络进行时序建模,并利用自注意力机制对卷积长短期记忆网络的神经元权重进行重新分配,提高了卷积长短期记忆网络对输入特征不确定性的适应能力。在此基础上,明确了自注意力机制和自适应噪声完全集合经验模态分解、加权排列熵在风电功率预测中的作用机制,以及风电功率信号包含的重要物理信息,证明了自适应噪声完全集合经验模态分解、加权排列熵以及自注意力机制在风电功率信号模态分解和长短期记忆网络隐层输出权重分配中的有效性。

基于动态双种群的黏菌和花粉混合算法

针对单一启发式算法易受自身原理导致的全局和局部搜索不平衡的问题,提出了一种基于动态双种群的黏菌和花粉混合算法HASMFP。首先,通过种群个体和当前最优个体之间的距离,将种群动态划分为黏菌子种群和花粉子种群分别进行搜索,以更有效地平衡算法的探索能力和开发能力;其次,对全局搜索融入相似度与适应度的综合排序感知机制来提高黏菌子种群的多样性,以帮助黏菌算法跳出局部最优;最后,在标准花粉算法的全局搜索中加入动态权重和恒定收缩系数,并对局部搜索设计了精英引导项来提高算法的收敛速度和搜索精度。选用CEC2017测试集中的12个函数作为实验测试集,将HASMFP与ISMA、DTSMA、HLFPA、SCFPA和tMFPA五种改进算法进行性能测试对比。还对HASMFP的各个改进策略进行消融实验,实验表明在综合改进策略的共同作用下,HASMFP的优化性能排名第一。基于实验结果的Friedman检验表明,HASMFP能够获取最优的性能。

基于犹豫模糊混合云模型的航空装备保障系统效能评估

针对航空装备保障系统效能评估过程存在定性指标与定量指标共存、评价信息具有不确定性等问题,提出了一种基于犹豫模糊混合云模型的效能评估方法。根据语言变量的转换与正态云模型之间的混合规则,将犹豫模糊语言术语集转换为梯形云模型,并根据云模型的数字特征,提出了犹豫模糊语言术语集的犹豫度函数与标度函数,实现了对定性指标的量化,兼顾了评价信息的犹豫性、模糊性和随机性。利用区间层次分析法结合灰狼优化算法对效能评估指标体系进行权重赋值,确定系统效能等级。通过应用实例与方法比较表明:犹豫模糊混合云模型不仅评估结果可靠,而且能够识别系统的运行状态,发现系统建设的短板,为提高系统效能提供重要的参考意见,更具有实用性。

基于多气象类型加权和改进高斯混合模型的光伏出力超短期概率预测

提出一种多气象类型加权和改进高斯混合模型,可实现对光伏出力提前15 min的超短期概率预测。首先,依据气象特征将历史数据划分为若干气象类型,然后,构建改进高斯混合模型获得每个气象类型出力数据的概率分布,其次,构建隶属度函数量化待预测时刻气象特征对于各气象类型的相似程度,最后,根据隶属度对各气象类型的概率分布加权。以实际光伏电站数据进行算例分析,结果表明相较于单一气象类型,多气象类型加权模型的MAPE平均减少16.87%,ACD平均提升10.45%,AW平均下降2.49%。

基于多视图与注意力机制的睡眠脑电分期

为了更全面地对睡眠脑电进行特征提取,提出一种基于多视图与注意力机制的睡眠脑电分期方法。首先针对原始睡眠脑电信号构造时域和时频域两类视图数据;然后设计融合注意力机制的混合神经网络对多视图数据进行表征学习;接着通过双向长短时记忆(bi-directional long short-term memory,BiLSTM)网络进一步学习睡眠阶段之间的转换规则;最后使用Softmax函数进行睡眠分期,并利用类别加权损失函数解决睡眠数据类别不均衡的问题。实验使用Sleep-EDF数据库中前20名受试者的单通道脑电信号并采用20折交叉验证对模型进行性能评估,睡眠分期准确率达到83.7%,宏平均F_(1)值达到79.0%,Cohen′s Kappa系数达到0.78。与现有方法相比,算法性能提升明显,证明了所提方法的有效性。

足月低出生体重儿12个月内发育状况及其影响因素分析

目的探讨分析足月低出生体重儿12个月内的发育状况及其影响因素。方法选取2021年2月—2022年9月出生的足月低出生体重儿80例作为研究对象,对其进行追踪随访12个月,观察其生长发育情况,并分析其影响因素。结果80例足月低出生体重儿3个月的平均体重、身长、头围分别为(5.57±0.83)kg、(58.86±3.49)cm、(38.70±2.59)cm;6个月的平均体重、身长、头围分别为(7.18±1.04)kg、(65.61±5.38)cm、(42.12±2.09)cm;12个月的平均体重、身长、头围分别为(8.66±1.75)kg、(73.44±4.86)cm、(44.58±1.79)cm;其中发育正常59例,发育落后21例。Logistic多因素回归分析结果显示,出生时体重、出生时身长、父母身高、喂养方式、营养素情况、母亲文化程度、有无疾病是影响足月低出生体重儿生长发育的主要因素(P<0.05)。结论足月低出生体重儿出生后12个月为追赶生长关键期,随月龄增长,足月低出生体重儿与正常出生体重儿的差异可逐渐缩小,但是多数追赶生长不太理想,在喂养过程中需注意合理喂养,加强随访追踪干预,加强家庭知识宣教,尽量减少诸多因素对其生长发育的影响。

基于考虑误差修正的非线性自适应权重组合模型的光伏发电功率预测

为了提高光伏电站光伏发电功率预测精度,解决极限梯度提升模型、长短期记忆模型2种传统单一模型及传统组合模型极限梯度提升-长短期记忆模型的光伏发电功率预测结果滞后、预测效果易突变、预测误差较大、线性拟合性较差等不足,基于极限梯度提升算法、长短期记忆算法和线性自适应权重,提出一种考虑误差修正的非线性自适应权重极限梯度提升-长短期记忆模型进行光伏发电功率预测;分别使用极限梯度提升算法和长短期记忆算法训练得到2种单一模型,将2种单一模型的初步预测值和真实值组成新的训练数据集,利用神经网络算法训练所提出的模型,对2种单一模型的初步预测值分配自适应权重系数,并根据训练时所提出模型的预测值大小分段统计预测误差的分布,预测时根据所提出模型的预测值在预测结果的基础上累加误差均值从而进行误差修正,进一步提高所提出模型的预测精度;利用Python语言分别对所提出的模型、传统组合模型和2种传统单一模型在晴天、阴天和雨天的光伏发电功率预测性能进行仿真。结果表明:与极限梯度提升-长短期记忆模型、极限梯度提升模型、长短期记忆模型相比,所提出模型的均方根误差分别减小28.57%、 39.39%、 49.79%,平均绝对误差分别减小44.25%、 53.33%、 64.8%,决定系数分别增大1.43%、 2.38%、 3.34%,所提出的模型更有效地减小了传统单一模型的光伏发电功率预测误差,优化了传统组合模型的权重系数;3种天气条件下所提出模型的光伏发电功率预测误差相对最小且稳健性最强,验证了所提出模型的有效性。

妊娠糖尿病患者孕期体质量变化与母婴不良结局的关系研究

目的探讨妊娠糖尿病(GDM)患者孕期体质量变化与母婴不良结局的关系。方法选取2020年1月至2021年12月监利市中医医院产科收治的80例GDM患者作为研究对象,根据体重指数(BMI)不同分为体质量正常组(n=32)、超重组(n=28)及肥胖组(n=20)。比较3组临床资料,采用Spearman相关性分析GDM患者孕期体质量变化与不良母婴结局的关系,采用多因素Logistic回归分析孕期体质量变化的影响因素。结果3组孕前体质量、孕期血糖控制效果、甘油三酯(TG)、低密度脂蛋白胆固醇(LDL-C)、孕期是否接受过饮食及运动干预比较差异有统计学意义(P<0.05);3组年龄、学历、孕次、孕龄、高密度脂蛋白胆固醇(HDL-C)、稳态模型评估-胰岛素抵抗指数(HOMA-IR)比较差异无统计学意义。体质量正常组巨大儿发生率、胎儿宫内窒息发生率、产妇出血率均低于超重组及肥胖组,差异有统计学意义(P<0.05)。Spearman相关性分析结果显示,GDM患者孕期体质量变化与足月巨大儿发生率呈正相关(r>0,P<0.05),而与胎儿宫内窒息发生率、产妇出血率无明显相关性。多因素Logistic回归分析结果显示,孕前体质量高、孕期血糖控制效果不佳、TG高水平、LDL-C高水平是GDM患者孕期体质量增长的危险因素(P<0.05),而孕期进行饮食及运动干预是GDM产妇孕期体质量增加的保护因素(P<0.05)。结论GDM患者孕期肥胖与足月巨大儿发生率呈正相关,孕期加强饮食及运动干预有助于控制患者体质量增长,减少不良母婴结局。

基于TCN-BiLSTM-AM的居民住宅短期电力负荷预测

针对当前住宅短期电力负荷预测模型存在预测精度低和特征提取困难等问题,提出一种基于TCN-BiLSTMAM的住宅电力负荷预测模型。该模型主要由TCN模型和引入注意力机制层改进的BiLSTM模型组成。首先,通过在历史数据中使用负荷曲线技术计算特征变量的输入权重,以提高数据输入的准确度和关联性;然后,采用权重匹配的方法将数据序列化输入到TCN模型进行采样训练,提取更多不同时间尺度的特征并加快训练速度,同时,构建改进的BiLSTM模型,引入AM层以提高BiLSTM网络结构的运算速度和处理长序列数据的能力,从而提高模型的泛化能力和运算速度;接着,通过对训练好的TCN模型和改进的BiLSTM模型进行加权输出初始预测值,并利用遗传算法对预测值与真实值的偏差进行偏置寻优,得到优化权重并输出最终预测结果。最后,在同一公开数据集上与RNN、LSTM、BiLSTM和TCN等模型进行对比验证,结果表明,相比较其中较好的模型,文中提出的TCN-BiLSTM-AM模型在MAE和RMSE上分别降低了40.43%和35.59%,同时R2指标为0.9957,具有更高的预测精度和更好的稳定性。