Predicting Users’ Latent Suicidal Risk in Social Media: An Ensemble Model Based on Social Network Relationships

Suicide has become a critical concern,necessitating the development of effective preventative strategies.Social media platforms offer a valuable resource for identifying signs of suicidal ideation.Despite progress in detecting suicidal ideation on social media,accurately identifying individuals who express suicidal thoughts less openly or infrequently poses a significant challenge.To tackle this,we have developed a dataset focused on Chinese suicide narratives from Weibo’s Tree Hole feature and introduced an ensemble model named Text Convolutional Neural Network based on Social Network relationships(TCNN-SN).This model enhances predictive performance by leveraging social network relationship features and applying correction factors within a weighted linear fusion framework.It is specifically designed to identify key individuals who can help uncover hidden suicidal users and clusters.Our model,assessed using the bespoke dataset and benchmarked against alternative classification approaches,demonstrates superior accuracy,F1-score and AUC metrics,achieving 88.57%,88.75%and 94.25%,respectively,outperforming traditional TextCNN models by 12.18%,10.84%and 10.85%.We assert that our methodology offers a significant advancement in the predictive identification of individuals at risk,thereby contributing to the prevention and reduction of suicide incidences.

Strategies to improve genomic predictions for 35 duck carcass traits in an F2 population

Background Carcass traits are crucial for broiler ducks,but carcass traits can only be measured postmortem.Genomic selection(GS)is an effective approach in animal breeding to improve selection and reduce costs.However,the performance of genomic prediction in duck carcass traits remains largely unknown.Results In this study,we estimated the genetic parameters,performed GS using different models and marker densi-ties,and compared the estimation performance between GS and conventional BLUP on 35 carcass traits in an F2 population of ducks.Most of the cut weight traits and intestine length traits were estimated to be high and moder-ate heritabilities,respectively,while the heritabilities of percentage slaughter traits were dynamic.The reliability of genome prediction using GBLUP increased by an average of 0.06 compared to the conventional BLUP method.The Permutation studies revealed that 50K markers had achieved ideal prediction reliability,while 3K markers still achieved 90.7%predictive capability would further reduce the cost for duck carcass traits.The genomic relationship matrix nor-malized by our true variance method instead of the widely used 2pi(1-pi)could achieve an increase in prediction reliability in most traits.We detected most of the bayesian models had a better performance,especially for BayesN.Compared to GBLUP,BayesN can further improve the predictive reliability with an average of 0.06 for duck carcass traits.Conclusion This study demonstrates genomic selection for duck carcass traits is promising.The genomic prediction can be further improved by modifying the genomic relationship matrix using our proposed true variance method and several Bayesian models.Permutation study provides a theoretical basis for the fact that low-density arrays can be used to reduce genotype costs in duck genome selection.

Tensile Properties and Prediction Model of Recombinant Bamboo at Different Temperatures

The destruction of recombinant bamboo depends on many factors,and the complex ambient temperature is an important factor affecting its basic mechanical properties.To investigate the failure mechanism and stress–strain relationship of recombinant bamboo at different temperatures,eighteen tensile specimens of recombinant bamboo were tested.The results showed that with increasing ambient temperature,the typical failure modes of recombinant bamboo were flush fracture,toothed failure,and serrated failure.The ultimate tensile strength,ultimate strain and elastic modulus of recombinant bamboo decreased with increasing temperature,and the ultimate tensile stress decreased from 154.07 to 96.55 MPa,a decrease of 37.33%,and the ultimate strain decreased from 0.011 to 0.008,a decrease of 26.57%.Based on the Ramberg-Osgood model and the pseudo‒elastic design method,a predictive model was established for the tensile stress–strain relationship of recombinant bamboo considering the temperature level.The model can accurately evaluate the tensile stress–strain relationship of recombinant bamboo under different temperature conditions.

Virtual Machine Consolidation with Multi-Step Prediction and Affinity-Aware Technique for Energy-Efficient Cloud Data Centers

Virtual machine(VM)consolidation is an effective way to improve resource utilization and reduce energy consumption in cloud data centers.Most existing studies have considered VM consolidation as a bin-packing problem,but the current schemes commonly ignore the long-term relationship between VMs and hosts.In addition,there is a lack of long-term consideration for resource optimization in the VM consolidation,which results in unnecessary VM migration and increased energy consumption.To address these limitations,a VM consolidation method based on multi-step prediction and affinity-aware technique for energy-efficient cloud data centers(MPaAF-VMC)is proposed.The proposed method uses an improved linear regression prediction algorithm to predict the next-moment resource utilization of hosts and VMs,and obtains the stage demand of resources in the future period through multi-step prediction,which is realized by iterative prediction.Then,based on the multi-step prediction,an affinity model between the VM and host is designed using the first-order correlation coefficient and Euclidean distance.During the VM consolidation,the affinity value is used to select the migration VM and placement host.The proposed method is compared with the existing consolidation algorithms on the PlanetLab and Google cluster real workload data using the CloudSim simulation platform.Experimental results show that the proposed method can achieve significant improvement in reducing energy consumption,VM migration costs,and service level agreement(SLA)violations.

GIS-Based Urbanization Prediction Model Considering Neighborhood Relationship of the Unit of the “Block” in the Outskirts of Metropolitan Area

On the outskirts of the metropolitan areas in Japan, the rapid development of urban areas and the improvement in transportation networks have brought various land use problems in their wake, including urban diffusion and the phenomenon of urban sprawl. There is a strong need for accurate predictions of land-use change and future urbanization, as well as investigation of the appropriateness of present land use controls and the land use controls that will be required in the future. This study took as its object the outskirts of the Keihanshin (Kyoto-Osaka-Kobe) Metropolitan Area, the second largest conurbation in Japan after the Tokyo Metropolitan Area, and used the digital maps and spatial analysis offered by GIS. It aimed to: 1) describe the characteristics of land use controls, land use and urbanization;2) develop an urbanization prediction model that considers the neighboring relationship of neighboring areas on a 100 m mesh unit;3) apply this model to the study area and verify its validity regarding the conditions of present land use;4) compare urbanization prediction results by this model with the present land use controls;and 5) make predictions for future urbanization and propose remedial measures for future land use controls.

A prediction model for horizontal run-out distance of landslides triggered by Wenchuan earthquake

The peak ground acceleration （PGA）, the volume of a sliding mass V, the height of a mountain HL and the slope angle θ of a mountain are four important parameters affecting the horizontal run-out distance of a landslide L. Correlations among them are studied statistically based on field investigations from 67 landslides triggered by the ground shaking and other factors during the Wenchuan earthquake, and then a prediction model for horizontal run-out distance L is developed in this study. This model gives due consideration to the implications of the above four parameters on the horizontal run-out distance L and the validity of the model is verified by the Donghekou and Magong Woqian landslides. At the same time, the advantages of the model are shown by comparing it with two other common prediction methods. The major findings drawn from the analyses and comparisons are： （1） an exponential relationship exists between L and log V, L and log HL, L and log PGA separately, but a negative exponential relationship exists between L and log tan0, which agrees with the statistical results; and （2） according to the analysis results of the relative relationship between the height of a mountain （H） and the place where the landslides occur, the probabilities at distances of2H/3-H, H/3-2H/3, and O-H/3 are 70.8%, 15.4%, and 13.8%, respectively, revealing that most landslides occurred at a distance of H/2-H. This prediction model can provide an effective technical support for the prevention and mitigation of landslide hazards.

The application of neural networks to comprehensive prediction by seismology prediction method

BP neural networks is used to mid-term earthquake prediction in this paper. Some usual prediction parameters of seismology are used as the import units of neural networks. And the export units of neural networks is called as the character parameter W_0 describing enhancement of seismicity. We applied this method to space scanning of North China. The result shows that the mid-term anomalous zone of W_0-value usually appeared obviously around the future epicenter 1~3 years before earthquake. It is effective to mid-term prediction.

Prediction of Gas Chromatographic Retention Indices of Organophosphates by DFT and VSMP Method

Polychlorinated dibenzothiophenes（PCDTs） are a group of important persistent organic pollutants.In the present study,geometrical optimization and electrostatic potential calculations have been performed for all 135 PCDTs congeners at the B3LYP/6-31G＊ level of theory.By means of the VSMP（variable selection and modeling based on prediction） program,one optimal descriptor（molecular polarizability,α） was selected to develop a QSRR model for the prediction of gas chromatographic retention indices（GC-RI） of PCDTs.The estimated correlation coefficients（r2） and LOO-validated correlation coefficients（q2）,all more than 0.99,were built by multiple linear regression,which shows a good estimation ability and stability of the models.A prediction power for the external samples was validated by the model built from the training set with 17 polychlorinated dibenzothiophenes.

Solubility study of hydrogen in direct coal liquefaction solvent based on quantitative structure–property relationships model

Direct coal liquefaction(DCL)is an important and effective method of converting coal into high-valueadded chemicals and fuel oil.In DCL,heating the direct coal liquefaction solvent(DCLS)from low to high temperature and pre-hydrogenation of the DCLS are critical steps.Therefore,studying the dissolution of hydrogen in DCLS under liquefaction conditions gains importance.However,it is difficult to precisely determine hydrogen solubility only by experiments,especially under the actual DCL conditions.To address this issue,we developed a prediction model of hydrogen solubility in a single solvent based on the machine-learning quantitative structure–property relationship(ML-QSPR)methods.The results showed that the squared correlation coefficient R^(2)=0.92 and root mean square error RMSE=0.095,indicating the model’s good statistical performance.The external validation of the model also reveals excellent accuracy and predictive ability.Molecular polarization(a)is the main factor affecting the dissolution of hydrogen in DCLS.The hydrogen solubility in acyclic alkanes increases with increasing carbon number.Whereas in polycyclic aromatics,it decreases with increasing ring number,and in hydrogenated aromatics,it increases with hydrogenation degree.This work provides a new reference for the selection and proportioning of DCLS,i.e.,a solvent with higher hydrogen solubility can be added to provide active hydrogen for the reaction and thus reduce the hydrogen pressure.Besides,it brings important insight into the theoretical significance and practical value of the DCL.

基于多序列隐关系的时序事件预测

时序事件预测是指基于历史事件预测下一个事件,事件包括时间和类型两个属性。当前主要工作集中在单方面(事件时间或事件类型)的预测,但这无法回答“何时发生何事”这类更精细的问题。此类问题的挑战主要是事件类型非常多样,而行为往往高度稀疏,给预测带来极大困难;需要预测的事件时间和事件类型分属两个域,如何把这两个域的信息加以融合并形成互补也是一个挑战。针对上述挑战,从融合多序列隐信息的角度探索了一种解决方法。基于某些事件序列之间具有模式相似性这一观察,提出建模事件序列的隐关系图,利用邻居序列的信息解决行为稀疏性的问题;通过合理设计神经网络模块,将事件的时间域和类型域的信息映射到共同的抽象空间,解决事件时间和事件类型信息的融合建模问题。通过在多个真实数据集上进行了大量实验,实验结果印证了多序列深度时序模型优于现有的一系列基准模型。

基于BP神经网络的胶结砂砾石应力-应变关系预测

在前期宏观试验基础之上,采用离散元模拟和BP神经网络相结合的方法获取不同胶凝材料掺量和围压下胶结砂砾石的应力-应变关系。根据前期胶凝材料掺量分别为20、40、60、80、100 kg/m3的胶结砂砾石三轴排水剪切试验结果,开展离散元数值模拟。以试验数据为学习样本,开展BP神经网络模型训练,预测胶凝材料掺量分别为30、50、70、90 kg/m3的胶结砂砾石应力-应变关系,并将预测结果和离散元模拟结果进行对比。研究结果表明,BP神经网络能够实现胶结砂砾石应力-应变关系的预测,并在较低围压下具有较好的精度。

基于STE-TCN的中短期电力负荷预测

目的 针对传统电力负荷预测模型对长序列预测精度低的问题,提出一种结合跳级卷积连接与时间编码网络的新型时序卷积神经网络(TCN)模型——STE-TCN模型。方法 首先对TCN模型加入跨周期的膨胀卷积通道(Skip-convolution)提取电力数据周期信息;再进行特征融合得到Skip-TCN网络,使网络抓取周期规律,增加信息利用长度;最后设计日期编码网络(Time encoding network)捕捉生活周期和季节性特征,与Skip-TCN进行特征融合得到STE-TCN模型,实现对电力负荷数据长序列预测。结果 实验表明:在与TCN模型和传统时序网络的对比下,Skip-TCN的预测精度均有提升,在预测长度更长的测试上提升尤为明显。结论 实验结果验证了通过对更长跨度时序关系的捕捉,STE-TCN网络改进方法有效提升了对长序列电力负荷的预测精度。

完全图高阶关系驱动的链接预测

图卷积网络(graph convolutional network,GCN)因其在处理图数据方面的独特优势而被广泛应用于推荐系统中,它通过利用图中节点之间的依赖关系传播节点属性信息,极大地提高了节点表示的准确度从而提升推荐性能.然而现有基于GCN的推荐方法仍因过平滑问题而难以进行更深层的建模,从而限制了用户与项目间高阶关系的表达.为此,提出了一种基于项目间关系的完全图高阶关系驱动的链接预测(link prediction driven by high-order relations in complete graph,LinkCG)方法.LinkCG通过用户-项目交互图与项目间隐式关联关系全局图组成的异构图预测用户到项目的链接,跳过了中间的用户节点直接利用完全图建模每个用户历史交互的项目间的局部隐式关联关系,获得项目间的高阶关系从而缓解数据稀疏性问题;此外,不同于基于节点嵌入的推荐方法,LinkCG通过赋予项目间的链接权重来表示项目间关系的紧密程度,并根据紧密程度进行链接预测,优化了模型的训练过程.在3个公开数据集上的实验结果表明,LinkCG作为只包含2个超参数的非深度学习模型,与一些先进的基于深度学习的基线方法相比提供了更好的性能.在社交关系数据上的应用进一步表明LinkCG能够从用户历史交互项目中获取足够丰富的用户偏好信息.

定量有害结局路径的构建方法及在毒理学中的应用研究进展

美国国家研究咨询委员会提出的“21世纪毒性测试——愿景与策略”对毒性评价和风险评估提出了新的要求和愿景,推动了新一代毒性测试方法和新一代风险评估方法学的发展。有害结局路径(AOP)作为一种先进的方法,具有较大应用潜力,近年来逐渐成为毒理学家关注的重点和新兴研究热点。定量AOP(qAOP)是在AOP理论上发展的新策略,其将定性AOP作为初始概念模型,通过数理模型描述剂量-响应和(或)响应-响应关系,提供了一种可定量预测体内毒性和风险的策略。本文就qAOP的概念和进展进行综述,介绍了2种构建qAOP的常用方法,贝叶斯网络模型和回归模型,并以实际案例展示qAOP在毒理学中的不同应用方向。

基于主成分分析淮北五道沟日尺度地温对气象要素的响应

为探讨淮北地区地温对气象要素的响应关系,对未来地温进行预测,利用五道沟水文实验站1986-2021年0-320cm深度土壤内9个土层地温和8个逐日气象要素(平均气温、水面蒸发、日照时数、降水量、1.5m高度平均风速、平均空气相对湿度、绝对湿度和水汽压力差)实测资料,分析地温与气象要素的相关关系,利用主成分分析法对8个气象要素进行数据降维,以提取出的主成分作为输入构建PCA-BP地温模拟模型,并与仅使用平均气温作为输入的BP神经网络模型进行对比分析。结果表明:(1)除平均气温外,地温还受水面蒸发、绝对湿度、水汽压力差等气象要素影响,气象要素对地温的影响随土壤深度的增加而减弱。深层地温对气象要素的响应表现出时滞性特点,滞后时间随土层深度增加而增加。(2)0-80cm处PCA-BP地温模型在预测未来1、3、7d后地温时精度较高,R^(2)(决定系数)均高于0.79。由于地温对气象要素响应的滞后性,预测天数增加时,PCA-BP地温模型对深层地温的预测精度提高。(3)引入新的气象要素作为输入,可提高地温预测模型的精度;与仅使用平均气温作为输入的BP模型对比,PCA-BP模型在预测未来3d和7d后地温表现更优。

滇中引水工程水源区水沙关系及悬移质含沙量预测方法研究

滇中引水工程水源状况对供水至关重要,为全面了解工程水源区的水量和悬移质含沙量特性,分析悬移质含沙量未来情势,依据水源区石鼓水文站资料系列,基于水文随机分析、小波分析、Copula函数等方法研究石鼓站水与沙的变化趋势及周期规律、联合分布及相依变化情况,在此基础上提出一种悬移质含沙量模拟及预测模型。研究结果表明,石鼓站流量和悬移质含沙量的年际、年内分配不均匀;水沙组合的同现重现期最大、联合重现期次之、条件重现期最小,同步遭遇概率远大于异步遭遇概率;未来丰、平、枯水情景下的多年平均悬移质含沙量分别为0.639、0.597、0.562 kg/m^(3),较1955~2019年实测的0.586 kg/m^(3)分别增长了9.0%、1.9%、-4.1%,3种情景下悬移质含沙量均呈上升发展趋势。研究成果从水沙分析角度为滇中引水工程供水调度、含沙量预估等提供了决策依据。

基于客户-客服沟通文本信息的客户流失研究

为探索客服与客户之间的交流是否会影响客户流失,以某互联网素质类在线教培公司的客户-客服对话数据为基础,通过文本分析方法挖掘对话中蕴含的各类主题以及情感状态,并结合逻辑回归模型进一步探索对话内容对客户流失的影响程度。研究表明:①互动过程中客服的礼貌用语可以减少客户流失;②互动过程中客户的情感态度和客户流失具有非线性关系;③客服与客户紧扣业务的积极互动会提升客户的续费率,而系统预设生成的群发消息无法引起客户的回应,会导致客户续费意愿的下降。

一种基于强化学习的软件安全实体关系预测方法

为改善现有基于翻译的软件安全知识图谱实体关系预测方法不具备可解释性,而基于路径推理的方法准确性不高的现状,本研究提出一种基于强化学习的预测方法 .该方法首先分别使用TuckER模型和SBERT模型将软件安全知识图谱的结构信息和描述信息表示为低维度向量,接着将实体关系预测过程建模为强化学习过程,将TuckER模型计算得到的得分引入强化学习的奖励函数,并且使用输入的实体关系向量训练强化学习的策略网络,最后使用波束搜索得到答案实体的排名列表和与之对应的推理路径.实验结果表明,该方法给出了所有预测结果相应的关系路径,在链接预测实验中hit@5为0.426,hit@10为0.797,MRR为0.672,在事实预测实验中准确率为0.802,精确率为0.916,在准确性方面与同类实体关系预测模型相比具有不同程度的提升,并且通过进行可解释性分析实验,验证了该方法所具备的可解释性.

基于权值对三角关系中结构平衡的预测研究

[研究目的]随着信息化社会的不断发展变化,开源情报的重要性日益突出。通过社交网络中的公开信息对实体间的关系演化进行预测,为使用者提供决策支持。[研究方法]利用微博文本数据挖掘技术,对一年内的官博文本进行收集和清洗,再运用共现次数及链路赋予数值等方法,构建社会关系网络,以预测结构不平衡的社交网络。[研究结论]研究结果表明,该方法在一定参数约束下能够生成符合预期的社会关系网络,并有效预测在社会关系网络中结构不平衡状态下可能出现的变化。

基于知识图谱推理的工控漏洞利用关系预测方法

工业控制系统漏洞数量呈快速增长态势,人工分析漏洞利用需要花费的时间与经济成本不断增加,当前推理方法存在信息利用不充分、可解释性差等缺陷.针对上述问题,提出了一种基于知识图谱推理的工控漏洞利用关系预测方法.该方法首先使用路径筛选算法约简漏洞利用路径,然后通过关键关系路径聚合获取路径信息,通过邻居关系信息融合获取邻居信息,最终预测漏洞利用关系.基于安全知识数据与工控场景数据构建了一个包含57333个实体的工控安全知识图谱,进行漏洞利用关系预测实验.结果表明,提出的方法预测准确率达到99.0%,可以辅助工控漏洞利用预测.