Development of a new Cox model for predicting long-term survival in hepatitis cirrhosis patients underwent transjugular intrahepatic portosystemic shunts

BACKGROUND Transjugular intrahepatic portosystemic shunt(TIPS)placement is a procedure that can effectively treat complications of portal hypertension,such as variceal bleeding and refractory ascites.However,there have been no specific studies on predicting long-term survival after TIPS placement.AIM To establish a model to predict long-term survival in patients with hepatitis cirrhosis after TIPS.METHODS A retrospective analysis was conducted on a cohort of 224 patients who un-derwent TIPS implantation.Through univariate and multivariate Cox regression analyses,various factors were examined for their ability to predict survival at 6 years after TIPS.Consequently,a composite score was formulated,encompassing the indication,shunt reasonability,portal venous pressure gradient(PPG)after TIPS,percentage decrease in portal venous pressure(PVP),indocyanine green retention rate at 15 min(ICGR15)and total bilirubin(Tbil)level.Furthermore,the performance of the newly developed Cox(NDC)model was evaluated in an in-ternal validation cohort and compared with that of a series of existing models.RESULTS The indication(variceal bleeding or ascites),shunt reasonability(reasonable or unreasonable),ICGR15,post-operative PPG,percentage of PVP decrease and Tbil were found to be independent factors affecting long-term survival after TIPS placement.The NDC model incorporated these parameters and successfully identified patients at high risk,exhibiting a notably elevated mortality rate following the TIPS procedure,as observed in both the training and validation cohorts.Additionally,in terms of predicting the long-term survival rate,the performance of the NDC model was significantly better than that of the other four models[Child-Pugh,model for end-stage liver disease(MELD),MELD-sodium and the Freiburg index of post-TIPS survival].CONCLUSION The NDC model can accurately predict long-term survival after the TIPS procedure in patients with hepatitis cirrhosis,help identify high-risk patients and guide follow-up management after TIPS implantation.

Prediction on effectiveness of road sweeping for highway runoff pollution control

Binzhou section of Changshen highway was selected to study the effectiveness of road sweeping in decreasing the pollutant loads of highway runoff.With on-site continuous sampling the discharge rules of Cu Cd Pb and Zn are analyzed.The total and dissolved event mean concentrations of Cu Pb and Zn are calculated and the loads of heavy metals attached to particles sampled before and after rainfall are also studied.A test section on highway road was swept in different frequencies during a week and the amount of removed particles was measured.Based on the analysis of highway runoff and road sweeping a prediction equation is established to calculate the pollution control efficiency of the sweeping measure and the results indicate that the 1 time/week road sweeping method can remove 47.93% of dissolved Cu 46.87% of Pb and 44.21% of Zn.

The long-term prediction of the oil-contaminated water from the Sanchi collision in the East China Sea

The condensate and bunker oil leaked from the Sanchi collision would cause a persistent impact on marine ecosystems in the surrounding areas. The long-term prediction for the distribution of the oil-polluted water and the information for the most affected regions would provide valuable information for the oceanic environment protection and pollution assessment. Based on the operational forecast system developed by the First Institute of Oceanography, State Oceanic Administration, we precisely predicted the drifting path of the oil tanker Sanchi after its collision. Trajectories of virtual oil particles show that the oil leaked from the Sanchi after it sank is mainly transported to the northeastern part of the sink location, and quickly goes to the open ocean along with the Kuroshio. Risk probability analysis based on the outcomes from the operational forecast system for years 2009 to2017 shows that the most affected area is at the northeast of the sink location.

Long-term Traffic Volume Prediction Based on K-means Gaussian Interval Type-2 Fuzzy Sets

This paper uses Gaussian interval type-2 fuzzy se theory on historical traffic volume data processing to obtain a 24-hour prediction of traffic volume with high precision. A K-means clustering method is used in this paper to get 5 minutes traffic volume variation as input data for the Gaussian interval type-2 fuzzy sets which can reflect the distribution of historical traffic volume in one statistical period. Moreover, the cluster with the largest collection of data obtained by K-means clustering method is calculated to get the key parameters of type-2 fuzzy sets, mean and standard deviation of the Gaussian membership function.Using the range of data as the input of Gaussian interval type-2 fuzzy sets leads to the range of traffic volume forecasting output with the ability of describing the possible range of the traffic volume as well as the traffic volume prediction data with high accuracy. The simulation results show that the average relative error is reduced to 8% based on the combined K-means Gaussian interval type-2 fuzzy sets forecasting method. The fluctuation range in terms of an upper and a lower forecasting traffic volume completely envelopes the actual traffic volume and reproduces the fluctuation range of traffic flow.

Lithium-ion battery degradation trajectory early prediction with synthetic dataset and deep learning

Knowing the long-term degradation trajectory of Lithium-ion(Li-ion) battery in its early usage stage is critical for the maintenance of the battery energy storage system(BESS) in reality. Previous battery health diagnosis methods focus on capacity and state of health(SOH) estimation which can receive only the short-term health status of the cell. This paper proposes a novel degradation trajectory prediction method with synthetic dataset and deep learning, which enables to grasp the characterization of the cell's health at a very early stage of Li-ion battery usage. A transferred convolutional neural network(CNN) is chosen to finalize the early prediction target, and the polynomial function based synthetic dataset generation strategy is designed to reduce the costly data collection procedure in real application. In this thread, the proposed method needs one full lifespan data to predict the overall degradation trajectories of other cells. With only the full lifespan cycling data from 4 cells and 100 cycling data from each cell in experimental validation, the proposed method shows a good prediction accuracy on a dataset with more than 100 commercial Li-ion batteries.

New scheme of anticipating synchronization for arbitrary anticipation time and its application to long-term prediction of chaotic states

How to predict the dynamics of nonlinear chaotic systems is still a challenging subject with important real-life applications. The present paper deals with this important yet difficult problem via a new scheme of anticipating synchronization. A global, robust, analytical and delay-independent sufficient condition is obtained to guarantee the existence of anticipating synchronization manifold theoretically in the framework of the Krasovskii-Lyapunov theory. Different from ＇traditional techniques （or regimes）＇ proposed in the previous literature, the present scheme guarantees that the receiver system can synchronize with the future state of a transmitter system for an arbitrarily long anticipation time, which allows one to predict the dynamics of chaotic transmitter at any point of time if necessary. Also it is simple to implement in practice. A classical chaotic system is employed to demonstrate the application of the proposed scheme to the long-term prediction of chaotic states.

Long-term Prediction and Verification of Rainfall Based on the Seasonal Model

Using the seasonal cross-multiplication trend model, monthly precipitation of eight national basic weather stations of Shaanxi Province from 2005 to 2010 was predicted, and the forecast results were verified using the rainfall scoring rules of China Meteorological Administration. The verification results show that the average score of annual precipitation prediction in recent six years is higher than that made by a professional forecaster, so this model has a good prospect of application. Moreover, the level of making prediction is steady, and it can be widely used in long-term prediction of rainfall.

A Lightweight Temporal Convolutional Network for Human Motion Prediction

A lightweight multi-layer residual temporal convolutional network model(RTCN)is proposed to target the highly complex kinematics and temporal correlation of human motion.RTCN uses 1-D convolution to efficiently obtain the spatial structure information of human motion and extract the correlation in the time series of human motion.The residual structure is applied to the proposed network model to alleviate the problem of gradient disappearance in the deep network.Experiments on the Human 3.6M dataset demonstrate that the proposed method effectively reduces the errors of motion prediction compared with previous methods,especially of long-term prediction.

A Neuro-Based Software Fault Prediction with Box-Cox Power Transformation

Software fault prediction is one of the most fundamental but significant management techniques in software dependability assessment. In this paper we concern the software fault prediction using a multilayer-perceptron neural network, where the underlying software fault count data are transformed to the Gaussian data, by means of the well-known Box-Cox power transformation. More specially, we investigate the long-term behavior of software fault counts by the neural network, and perform the multi-stage look ahead prediction of the cumulative number of software faults detected in the future software testing. In numerical examples with two actual software fault data sets, we compare our neural network approach with the existing software reliability growth models based on nonhomogeneous Poisson process, in terms of predictive performance with average relative error, and show that the data transformation employed in this paper leads to an improvement in prediction accuracy.

National Prediction of Ambient Fine Particulates: 2000-2009

A large body of evidence links ambient fine particulates (PM2.5) to chronic disease. Efforts continue to be made to improve large scale estimation of this pollutant for within-urban environments and sparsely monitored areas. Still questions remain about modeling choices. The purpose of this study was to evaluate the performance of spatial only models in predicting national monthly exposure estimates of fine particulate matter at different time aggregations during the time period 2000-2009 for the contiguous United States. Additional goals were to evaluate the difference in prediction between federal reference monitors and non-reference monitors, assess regional differences, and compare with traditional methods. Using spatial generalized additive models (GAM), national models for fine particulate matter were developed, incorporating geographical information systems (GIS)-derived covariates and meteorological variables. Results were compared to nearest monitor and inverse distance weighting at different time aggregations and a comparison was made between the Federal Reference Method and all monitors. Cross-validation was used for model evaluation. Using all monitors, the cross-validated R2 was 0.76, 0.81, and 0.82 for monthly, 1 year, and 5-year aggregations, respectively. A small decrease in performance was observed when selecting Federal Reference monitors only (R2 = 0.73, 0.78, and 0.80 respectively). For Inverse distance weighting (IDW), there was a significantly larger decrease in R2 (0.68, 0.71, and 0.73, respectively). The spatial GAM showed the weakest performance for the northwest region. In conclusion, National exposure estimates of fine particulates at different time aggregations can be significantly improved over traditional methods by using spatial GAMs that are relatively easy to produce. Furthermore, these models are comparable in performance to other national prediction models.

Quantitative electroencephalography in predicting on outcome of awakening in long-term unconscious patients after severe traumatic brain injury

Objective To explore quantitative electroencephalography in unconscious patients after severe traumatic brain injury (TBI) to predict awakening. Methods All cases were divided into two groups(the awake group 19 cases and the unfavourable prognosis group 22 cases).Two weeks after admission the original EEGs were preformed in 41 patients suffering from severe TBI with duration of disturbance of

Long-Term Outcomes after Coronary Artery Bypass Grafting with Risk Stratification

Background: Risk stratification of long-term outcomes for patients undergoing Coronary artery bypass grafting has enormous potential clinical importance. Aim: To develop risk stratification models for predicting long-term outcomes following coronary artery bypass graft (CABG) surgery. Methods: We retrospectively revised the electronic medical records of 2330 patients who underwent adult Cardiac surgery between August 2016 and December 2022 at Madinah Cardiac Center, Saudi Arabia. Three hundred patients fulfilled the eligibility criteria of CABG operations with a complete follow-up period of at least 24 months, and data reporting. The collected data included patient demographics, comorbidities, laboratory data, pharmacotherapy, echocardiographic parameters, procedural details, postoperative data, in-hospital outcomes, and follow-up data. Our follow-up was depending on the clinical status (NYHA class), chest pain recurrence, medication dependence and echo follow-up. A univariate analysis was performed between each patient risk factor and the long-term outcome to determine the preoperative, operative, and postoperative factors significantly associated with each long-term outcome. Then a multivariable logistic regression analysis was performed with a stepwise, forward selection procedure. Significant (p < 0.05) risk factors were identified and were used as candidate variables in the development of a multivariable risk prediction model. Results: The incidence of all-cause mortality during hospital admission or follow-up period was 2.3%. Other long-term outcomes included all-cause recurrent hospitalization (9.8%), recurrent chest pain (2.4%), and the need for revascularization by using a stent in 5 (3.0%) patients. Thirteen (4.4%) patients suffered heart failure and they were on the maximum anti-failure medications. The model for predicting all-cause mortality included the preoperative EF ≤ 35% (AOR: 30.757, p = 0.061), the bypass time (AOR: 1.029, p = 0.003), and the duration of ventilation following the operation (AOR: 1.237, p = 0.021). The model for risk stratification of recurrent hospitalization comprised the preoperative EF ≤ 35% (AOR: 6.198, p p = 0.023), low postoperative cardiac output (AOR: 3.622, p = 0.007), and the development of postoperative atrial fibrillation (AOR: 2.787, p = 0.038). Low postoperative cardiac output was the only predictor that significantly contributed to recurrent chest pain (AOR: 11.66, p = 0.004). Finally, the model consisted of low postoperative cardiac output (AOR: 5.976, p < 0.001) and postoperative ventricular fibrillation (AOR: 4.216, p = 0.019) was significantly associated with an increased likelihood of the future need for revascularization using a stent. Conclusions: A risk prediction model was developed in a Saudi cohort for predicting all-cause mortality risk during both hospital admission and the follow-up period of at least 24 months after isolated CABG surgery. A set of models were also developed for predicting long-term risks of all-cause recurrent hospitalization, recurrent chest pain, heart failure, and the need for revascularization by using stents.

基于CEEMDAN-QPSO-BLS模型的径流预测研究

准确的径流预测是水资源优化配置和高效利用的前提,是制定防洪减灾决策的基础,然而受到人类活动、环境、气候等因素的影响,径流序列呈现出非线性、非稳态、多尺度变化的特点,这为径流的精准预测增加了难度。为提高径流预测的精准度和可信度,结合自适应噪声完备集合经验模态分解(Complete Ensemble Empirical Mode Decomposition with Adaptive Noise,CEEMDAN)方法,量子粒子群优化算法(Quantum Particle Swarm Optimization,QPSO)、宽度学习系统(Broad Learning System,BLS)模型,提出了一种基于CEEMDAN-QPSO-BLS组合式的径流预测模型。该组合模型首先使用CEEMDAN方法对原始径流信号进行分解,得到若干相对平稳的本征模态分量。其次利用QPSO算法对BLS模型的特征层节点组数、增强层节点组数和组内节点数进行寻优,得到最优的宽度学习网络拓扑结构,进而使用最优的QPSOBLS对多个稳态分量进行预测,并对预测分量进行重构,从而获得更高的预测精度。以黄河流域小浪底水库的日径流值为实验数据,将EMD-QPSO-BLS、QPSO-BLS作为CEEMDAN-QPSO-BLS的对比模型,并采用纳什效率系数(NSE)、均方根误差(RMSE)、平均绝对误差(MAE)和平均绝对百分比误差(MAPE)作为模型预测可信度和精准度的评价指标。实验表明,在预见期4天内,与QPSO-BLS、EMD-QPSO-BLS模型相比,CEEMDAN-QPSO-BLS的预测精准度分别提高了79.87%、19.80%,可信度分别提高了131.2%、10.98%,径流预测精度的提高,可为防洪抗旱保护人民生命财产和可持续发展提供决策支持。

基于GRU-CNN双网络输出构建BP模型的径流预测方法

提高径流预测精度是避免洪水灾害发生的重要手段,由于预测阶段并无已知有效样本,给预测工作带来难度,因此,提出以双网络输出为预测阶段提供数据参考,结合训练阶段双网络输出与真实值之间的关系,对预测阶段采用二次多变量建模实现径流预测。首先,构建GRU和CNN深度学习网络,同步输出2条径流预测序列;其次,在已知时段内,构建2条预测结果与实测值之间的多变量BP模型;最后,基于双网络输出预测值,通过确定的BP模型输出径流预测结果。经测试,该方法给预测时段提供了可靠的先验样本,高效学习了网络输出与真实值之间关系,预测精度显著提升。

径流序列相空间重构的水文学含义及应用

为确定径流序列相空间重构后的水文学含义并提高径流中长期预测精度,基于混沌理论进行径流序列相空间重构,并对径流影响因素与重构后相空间列向量进行相关性分析。在此基础上建立了混沌理论与人工神经网络耦合(Chaos-BPNN)的径流预测模型,并应用于黑河上游莺落峡水文站和正义峡水文站。结果表明:径流序列重构后相空间列向量具有明确的水文学含义;Chaos-BPNN径流预测模型仅需径流序列数据就可进行建模和预测,规避了径流预测过程中主控因素难以确定和不易量化的问题;黑河上游降水量、输沙量、水位和气温分别与重构后相空间的第1、3、6、7列具有较高的相关性,风速与任何一列都不相关,推测雪线高程、植被覆盖率以及土地利用类型等因素与第2、4、5列存在相关性;构建的Chaos-BPNN径流预测模型在黑河上游莺落峡水文站和正义峡水文站的径流预测精度均在86%以上。

基于多模型融合的中长期径流集成预测方法

中长期水文预报是流域水资源规划与合理配置的重要依据。为提高中长期径流预测精度,提出了一种基于多模型融合的水库中长期径流集成预测方法。该方法将ARMA、BP、LSTM、RF和SVR等5个异质预测模型进行融合,同时采用超参数优化方法确定各模型的最优参数。将其用于青海省龙羊峡水库的中长期径流预报中,结果表明,通过Stacking融合算法建立的集成预测模型相较于单一模型,取得了更高的预测精度(R2值由0.71提升至0.82)。此方法可为提升流域中长期径流预测精度提供一定参考。

基于多因子多模式集成的中长期径流预测模型

提高中长期径流预测精度对于水资源调度等具有重要意义和应用价值。基于国家气候中心的130项气候因子,采用皮尔逊相关系数、最大信息系数、方差增量指标筛选主要预测因子,建立基于DS(Dempster-Shafer)证据理论的多因子综合方法;采用随机森林、BP神经网络和贝叶斯网络等建立基于水文-气象因子遥相关的中长期径流预测模型,构建基于DS证据理论的预测结果集成模型。以三峡水库为对象开展实例研究,结果表明:引入遥相关因子能有效提高预测精度;基于DS证据理论的多因子综合方法能筛选出综合性更强、稳定性更优的因子,弥补单一筛选方法的不足;基于DS证据理论的多因子多模式集成方法在径流预测精度上优于单一方法单一模型,确定性系数提高到0.823,平均相对误差降低到23.2%。

宽度-深度融合时频分析的径流智能预测方法

为解决现有基于LSTM的径流预测模型易陷入局部最优的问题,提出了基于VMD-LSTMBLS(variational mode decomposition-LSTM-broad learning system)的径流预测模型。将宽度学习系统与LSTM结合,针对径流序列多噪音特点,采用时频分析方法中的变分模态分解,将径流时间序列的一维时域信号变换到二维时频平面,减少噪声对预测结果的影响。仿真结果表明:与基线模型及现有基于LSTM的径流预测模型相比,该模型的预测精度有较为明显的提高。

基于WPT-IDBO-RELM和WPT-IDMO-RELM模型的日径流预测

为提高日径流时间序列预测精度,改进正则化极限学习机(RELM)的预测性能,对比验证改进蜣螂优化(IDBO)算法和改进侏獴优化(IDMO)算法与其他算法的优化效果,提出了基于小波包变换(WPT)的WPT-IDBO-RELM和WPT-IDMO-RELM日径流时间序列预测模型。对云南省暮底河水库、马鹿塘电站入库日径流进行预测,结果表明WPT-IDBO-RELM和WPT-IDMO-RELM模型对暮底河水库日径流预测的平均绝对百分比误差分别为1.048%、1.015%,对马鹿塘电站日径流预测的平均绝对百分比误差分别为1.493%、1.478%,优于其他对比模型;IDBO、IDMO算法对标准测试函数和实例目标函数的寻优效果均优于其他对比算法,且IDBO、IDMO算法优化效果越好,RELM超参数越优,WPT-IDBO-RELM、WPT-IDMO-RELM模型预测精度越高;WPT可将日径流序列分解为分量更少、规律性更强的子序列分量,在提高预测精度的同时显著降低模型复杂度和计算规模。

基于物联网和GCNN-LSTM的河流水文预测方法

针对河流水文存在预测精度不高的问题,利用物联网技术设计了分布式的降雨和水文信息自动采集系统,并提出了一种基于图卷积神经网络和长短期记忆网络模型对河流水位和径流量进行预测的方法;首先通过分析确定了影响河流水文的主要因素,将流域范围内的降雨量信息组成网格化的二维图形矩阵;然后提出了GCNN-LSTM预测模型,将含有降雨信息的二维图形矩阵作为网络模型的输入,获取该流域内降雨与水文变化的时空分布特征;最后采用所提出的GCNN-LSTM预测模型对河南省周口市段颍河的历史水文数据进行训练,再利用训练后的网络对测试集数据进行预测,得到了较高精度的径流量和水位结果,径流量预测结果的RMSE、MAPE和MAE分别仅为17.09 m^(3)/s、1.68%和8.57 m^(3)/s,水位预测结果的RMSE、MAPE和MAE分别仅为0.32 m、0.65%和0.29 m,与其他几种预测方法相比表现出了优越性,对科学合理利用水资源和防洪减灾具有重要意义。