Search

Sea surface height (SSH) prediction is theoretically and practically significant for global and regional ocean-related research. Numerous studies have been conducted to acquire accurate prediction results. However, most investigations on SSH ignore the importance of data at each time step on the prediction, which limits the accuracy of the final prediction. Therefore, a deep learning model combined Long Short-Term Memory (LSTM) network and Attention mechanism is proposed in this letter. This model integrates attention mechanism in both of time and space dimensions into LSTM. For time dimension, it assigns reasonable weight for data at each time step. For space dimension, it groups the data points close to each other, let model concentrate on points in the same group and eliminates the impact from other points. Daily absolute dynamic topography (ADT) in the South China Sea from January 2010 to December 2017 is adopted to conduct experiments. The proposed model demonstrates reliable results, the root mean square error is 0.38 cm, the mean absolute error is 0.0031, and the correlation coefficient reaches up to 0.999. The results show that the deep learning method based on attention mechanism is reliable for SSH prediction with high performance.

Aims Surveillance imaging is often used to detect remodelling, a change in cardiac geometry, and/or function; however, there are limited data in patients with chronic heart failure (HF). We sought to characterize cardiac remodelling in patients with chronic HF and evaluate its association with outcome. Methods and results A prospective cohort of patients at risk for HF or with chronic HF underwent cardiac magnetic resonance (CMR) at baseline and 1 year. Ventricular function, volumes, mass, left atrial volume, global longitudinal strain, and myocardial scar were measured. The primary outcome was a composite of death or cardiovascular hospitalization up to 5 years from the 1 year scan. Cox regression was used to identify 1 year CMR predictors of outcome after adjusting for baseline risk. A total of 262 patients (median age 68 years, 57% males) including 96 at risk for HF, 97 with HF and preserved ejection fraction, and 69 with HF and reduced ejection fraction were included. In the patients with HF, 55 events were identified during follow‐up. After adjustment for baseline clinical risk, Cox proportion hazard regressions only identified 1 year change in left ventricular (LV) mass index as a CMR predictor of outcome, adjusted hazard ratio 1.21 (1.02, 1.44) per 10% increase, P = 0.031. Cardiac remodelling defined as a 1 year change in LV mass index ≥15% was observed in 35% of patients with HF. Patients with adverse remodelling of LV mass index had more events on Kaplan–Meier analyses compared to those with no remodelling, log‐rank P = 0.004 for overall cohort, P = 0.035 for heart failure with preserved ejection fraction and P = 0.035 for heart failure and reduced ejection fraction. Conclusions Cardiac remodelling is common during serial CMR assessment of patients with chronic HF. Change in LV mass predicted long‐term outcomes whereas change in left ventricular ejection fraction did not.

Wind speed accurate forecasting plays an important role in preserving the stability of offshore wind power. Most of current predictions are based on a single wind node. It is difficult for this methods to capture wind high dimensional features and latent spatio-temporal dependencies. In this paper, we propose a general graph optimization neural network specifically for multi-node offshore wind speed prediction named spatio-temporal correlation graph neural network. The proposed model firstly employs graph convolution, which performs Laplace transforms instead of the traditional convolution, to capture the potential spatial dependencies from the nodes' relationship and historical time series better. The channel-wise attention makes the original concentrated weights within a region constructed by adjacency matrix disperse to all input nodes, which distinguishes the nodes’ contributions and generates high dimensional spatial features. Long short-term memory is applied to extract temporal correlation from the high dimensional spatial features. The model has ability to excavate the full potential of spatial-temporal dependencies from multiple wind nodes to the utmost. Experiments select 120 wind nodes in the China sea for prediction. The results show that the proposed model can be very competitive with state-of-the-art methods and holds great performance on multi-node and multi-step wind speed forecasting.

Sustaining passengers’ loyalty towards bus service has been regarded as the key to secure its market share. However, passengers are not homogenous in terms of behavioral loyalty or attitudinal loyalty. Therefore, the determinants to retain loyalty may differ across each group of passengers. This study proposed a data-fusion based clustering method to segment passengers. An empirical study of Xiamen, China was given. Specifically, by fusing the smart card data and the survey data, the method measured passengers’ behavioral and attitudinal loyalty simultaneously. As a result, spurious-loyalty passengers were identified, who were characterized as behaviorally loyal but attitudinal disloyal; they held the most potential to shift to other transport modes. In order to retain the spurious-loyalty passengers, a novel model framework was built to investigate the determinants of driving them to continue using public transportation. The results indicated that the passengers’ satisfaction fully mediated between service value and attitudinal loyalty, as found in previous studies. Service value was unveiled to exert a significant and positive impact on predicting the passengers’ satisfaction and attitudinal loyalty. However, service value was not only determined by the passengers’ utilitarian value (perceived service quality), but also involved with their hedonic value (emotional dependence on public transportation which came from the passengers’ long-term involvement); hedonic value was demonstrated highly correlated with utilitarian value. In addition, we justified that utilitarian value could be separately measured by the quality of basic service attributes and the quality of value-added service attributes; they were revealed to exhibit non-linear and heterogeneous effects on the passengers’ satisfaction. Moreover, the index score of variables was employed to indicate the specific service attributes that needed to be improved to sustain the targeted passengers; practical suggestions to the specific direction of improvement were provided for transport sectors and public transport operators.

Causality is getting more and more attention, and its core idea is counterfactual and intervention. However, the current intervention model requires some prior knowledge, and lacks universality. The paper presents a novel solution called Search the Intervention Sample in Sparse Operation Space (SISSOS). SISSOS introduces variational inference and realizes intervention, that’s feature manipulation at the attribute level. SISSOS is for tabular data and uses sparse space to solve attribute coupling. SISSOS is applied to counterfactual and model interpretation in experiments. In the counterfactual experiment, the proposed solution was proven to find the correct causal effect without any prior knowledge. In the model interpretation experiment, a trained time series neural network with high accuracy was proved by the proposed solution to conform to prior knowledge. Compared with the previous method, the proposed method does not require prior knowledge and its intervention effect is better.

Objective To classify children with autism spectrum disorder (ASD) and typical development (TD) using short-term spontaneous hemodynamic fluctuations and to explore the abnormality of inferior frontal gyrus and temporal lobe in ASD. Methods 25 ASD children and 22 TD children were measured with functional near-infrared spectroscopy located on the inferior frontal gyrus and temporal lobe. To extract features used to classify ASD and TD, a multi-layer neural network was applied, combining with a three-layer convolutional neural network, a layer of long and short-term memory network (LSTM) and a layer of LSTM with Attention mechanism. In order to shorten the time of data collection and get more information from limited samples, a sliding window with 3.5 s width was utilized after comparisons, and numerous short (3.5 s) fNIRS time series were then obtained and used as the input of the multi-layer neural network. Results A good classification between ASD and TD was obtained with considerably high accuracy by using a multi-layer neural network in different brain regions, especially in the left temporal lobe, where sensitivity of 90.6% and specificity of 97.5% achieved. Conclusions The “CLAttention” multi-layer neural network has the potential to excavate more meaningful features to distinguish between ASD and TD. Moreover, the temporal lobe may be worth further study. Significance The findings in this study may have implications for rapid diagnosis of children with ASD and provide a new perspective for future medical diagnosis.

The study of sea surface temperature (SST) in coastal water is of great significance for navigation, aquaculture and military. Numerous studies have been conducted to predict this parameter in rece...

Sea surface temperature (SST) prediction plays an important role in ocean-related fields. It is challenging due to the nonlinear temporal dynamics with changing complex factors and the inherent difficulties in long-scale predictions. Conventional models often lack efficient information extraction and cannot meet the requirements of long-scale predictions. Therefore, the gate recurrent unit (GRU) encoder–decoder with SST codes and dynamic influence link (DIL), GRU encoder–decoder (GED), which considered both the static and dynamic influence, is proposed in this letter. Each SST code, capturing the static information more effectively, was computed by all hidden states of the encoder and was individually associated with each predicted SST. The DIL, capturing the dynamic influence, connected the SST code with the early predicted future SST for solving the long-scale dependence problem. GED was tested on the Bohai Sea SST data sets and South China Sea SST data sets and compared with full-connected long-short term memory (FC-LSTM) and support vector regression. The results demonstrated that GED outperformed others on different prediction scales and different prediction terms (daily, weekly, and monthly), especially in terms of long-scale and long-term predictions. In addition, attention relationships between historical and future SSTs were further explored, and there was a meaningful finding that each future daily mean SST of Bohai Sea most strongly correlated with the past 27th to 29th historical values.

Background Global longitudinal strain (GLS), most commonly measured at the endocardium, has been shown to be superior to left ventricular (LV) ejection fraction (LVEF) for the identification of systolic dysfunction and prediction of outcomes in heart failure (HF). We hypothesized that strains measured at different myocardial layers (endocardium = ENDO, epicardium = EPI, average = AVE) will have distinct diagnostic and predictive performance for patients with HF. Methods Layer-specific GLS, layer-specific global circumferential strain (GCS) and global radial strain (GRS) were evaluated by cardiovascular magnetic resonance imaging (CMR) feature tracking in the Alberta HEART study. A total of 453 subjects consisted of healthy controls (controls, n = 77), at-risk for HF (at-risk, n = 143), HF with preserved ejection fraction (HFpEF, n = 87), HF with mid-range ejection fraction (HFmrEF, n = 88) and HF with reduced ejection fraction (HFrEF, n = 58). For outcomes analysis, CMR-derived imaging parameters were adjusted with a base model that included age and N-terminal prohormone of b-type natriuretic peptide (NT-proBNP) to test their independent association with 5-year all-cause mortality. Results GLS_EPI distinguished all groups with preserved LVEF (controls − 16.5 ± 2.4% vs. at-risk − 15.5 ± 2.7% vs. HFpEF − 14.1 ± 3.0%, p Conclusion Global strains measured on endocardium, epicardium or averaged across the wall thickness are not equivalent for the identification of systolic dysfunction or outcomes prediction in HF. The endocardium-specific strains were shown to have poorest all-around performance. GLS_AVE and GRS were the only CMR parameters to be significantly associated with 5-year all-cause mortality in multivariable analysis. GLS_EPI and GRS, as well as the difference in endocardial and epicardial strains, were sensitive to systolic dysfunction among HF patients with normal LVEF (> 55%), in whom lower strains were associated with increased concentricity.

BackgroundThis study investigated accuracy and consistency of epicardial adipose tissue (EAT) quantification in chest computed tomography (CT) scans.Methods and resultsEAT volume was quantified semi-automatically using a standard Hounsfield unit threshold (-190U, -30) in three independent cohorts: (1) Cohort 1 (N= 30) consisted of paired 120 KV cardiac non-contrast CT (NCCT) and 120 KV chest NCCT; (2) Cohort 2 (N= 20) consisted of paired 120 KV cardiac NCCT and 100 KV chest NCCT; (3) Cohort 3 (N= 20) consisted of paired chest NCCT and chest contrast-enhanced CT (CECT) datasets. Images were reconstructed with the slice thicknesses of 1.25 mm and 5 mm in the chest CT datasets, and 3 mm in the cardiac NCCT datasets. In Cohort 1, the chest NCCT-1.25 mm EAT volume was similar to the cardiac NCCT EAT volume, whilst chest NCCT-5 mm underestimated the EAT volume by 7.0%. In Cohort 2, 100 KV chest NCCT-1.25mm and -5 mm EAT volumes were 9.7% and 6.4% larger than corresponding 120 KV cardiac NCCT EAT volumes. In Cohort 3, the chest CECT dataset underestimated EAT volumes by ∼25%, relative to chest NCCT datasets. All chest CT-derived EAT volumes were strongly correlated with their cardiac CT counterparts.ConclusionsThe chest NCCT-1.25 mm EAT volume with the 120 KV tube energy produced EAT volumes that are comparable to cardiac NCCT. All chest CT EAT volumes were strongly correlated with EAT volumes obtained from cardiac CT, if imaging protocol is consistently applied to all participants.

Background and Aims The current diagnostic criteria for acute kidney injury (AKI) predict the need for dialysis and early mortality, but are less useful to predict long-term outcomes. Acute kidney disease (AKD) defines patients with AKI or subacute loss of kidney function lasting for more than 7 days, which should predict better subsequent chronic kidney disease (CKD). The aim of this study was to investigate the risk factors and prognosis of AKD and to compare different types of acute/subacute renal impairment among Chinese inpatients. Method From a cohort of 450,000 patients consecutive admitted from June 1, 2012, to March 31, 2018 to five district hospitals, complete data were available from 71,041 inpatients. AKI and AKD were diagnosed based on the Acute Disease Quality Initiative Criteria 2017. Based on this diagnostic criterion of AKI and AKD, patients were classified as having (1) AKI Recover, if Scr back to baseline value within 7 days (renal impairment duration of less than 7 days or rapid recovery within 7 days), and (2) AKD with AKI, if a condition in which stage 1 or greater AKI was present ≥ 7 days after an AKI initiating event (continuous AKI progressing to AKD), (3) AKD without AKI, if Scr levels increased slowly but lasted more than 7 days (subacute AKD without meeting the AKI criterion). Results Of 71,041 inpatients, 16,098 (22.66%) patients developed AKI or AKD. 5,895 (8.30%) AKI patients recovered within 7 days (AKI Recover), 5,623 (7.91%) were followed by AKD and 4,580 (6.44%) patients developed AKD without AKI. Thus, AKI and AKD are frequent complications in Chinese inpatients (Fig 1). Compared to AKI recover or AKD without AKI, patients with AKI followed by AKD had higher hospital mortality (16.59% vs. 3.82% vs. 2.12%, P The AKI criterion yielded a higher risk for overall survival and a lower risk for de novo CKD than the AKD criterion, indicating that both criteria imply persistent kidney damage but that a rapid decline in excretory kidney function implies higher mortality risks while a persistent decline may rather result in de novo CKD (Fig 2). Meanwhile, these associations between different kidney injury criteria and outcomes had good generalizability and were constant across different genders, surgeries, and comorbidities (Fig 2). The AKD criterion was robustly associated with overall survival (area under the receiver operating characteristic curve (AUROC) 0.71) and de novo CKD (AUROC 0.71), while AKI criterion showed a relatively lower ability to fitting risk of overall survival (AUROC 0.65, P Conclusion (1) Adding AKD as a definition for renal failure lasting >7 days up to 90 days is of clinical importance in addition to the existing definitions for AKI and CKD. (2) These findings suggest research activities and clinical practice should also focus on AKD, which is far more accurate to predict subsequent de novo CKD.

Background and Aims Acute kidney injury (AKI) due to nephrectomy is associated with adverse prognosis. The study aimed to develop a nomogram based on preprocedural features for early prediction of AKI and to assess the prognosis in patients after nephrectomy. Method The study included a development cohort of 1111 patients and an additional validation cohort of 356 patients who underwent nephrectomy. Stepwise and logistic regression analyses were used to evaluate the association between predictors and AKI. Incorporating all independent predictors, a nomogram for postoperative AKI was developed and validated. Patients were followed up for 5 years to assess renal fuction, acute kidney disease (AKD), chronic kidney disease (CKD), hospital readmission and death were key prognostic outcomes we focused on. Results Radical nephrectomy (odds ratio (OR) = 3.57, p < 0.001), aspirin (OR = 1.79, p = 0.008), systolic blood pressure (OR = 1.41, p = 0.004), triglyceride (OR = 1.26, p = 0.024) were independent risk factors for postoperative AKI, while albumin (OR = 0.72, p = 0.031) was a protective factor for postoperative AKI. Patients with a higher estimated glomerular filtration rate (eGFR) (60-90 ml/min/1.73 m2, OR = 0.41, p = 0.004; ≥ 90 ml/min/1.73 m2, OR = 0.37, p < 0.001) were less prone to AKI than those with a lower eGFR (< 15 ml/min/1.73 m2). These predictors were all included in the final nomogram to calculate the probability of postoperative AKI. The incidence of AKD and CKD were 27.12% and 18.64% in AKI group, which were much higher than those in no AKI group (p < 0.001). Conclusion The nomogram had excellent predictive ability and might have significant clinical implications for the early detection of AKI in patients undergoing nephrectomy.

Natural resources depletion is gradually becoming a critical burden on the environmental and ecological balance, pushing the development of artificial aggregates forward. In order to curb the shortage issues of natural aggregates and minimize the destruction of land topography, alkali-activated aggregates (AAA) have gradually become a hot topic in recent years as a new application of alkali-activated materials (AAM) with the benefits of utilizing industrial by-products and waste materials. This article provides an overall review of the manufacturing process and engineering properties of two types of artificial aggregates, cold-bonded AAA (CB-AAA) and sintered AAA (ST-AAA). Their applications in concrete, besides the mechanical evaluation, durability performance, and leaching behavior, are summarized based on the existing research outcomes. Finally, the future perspectives and challenges of artificial aggregates development are also proposed.

The increase of scale and complexity of Internet big data presents unprecedented opportunities on Cyber-Physical Systems (CPS). The incompleteness and incredibility of Internet big data are challen...

Paclitaxel, a standard chemotherapeutic agent for several types of cancer, including ovarian, breast, and non-small-cell lung cancer, causes peripheral neuropathy as an adverse effect in 60-70% of the patients. The utility of combination therapy with paclitaxel and goshajinkigan, a traditional Japanese Kampo medicine, in managing paclitaxel-induced neuropathy during chemotherapy has been explored. Paclitaxel is predominantly metabolized in the liver by cytochrome P450 (CYP) 2C8 to produce 6α-hydroxypaclitaxel and by CYP3A4 to produce 3'-p-hydroxypaclitaxel. In this study, we evaluated the inhibitory or inducing effects of goshajinkigan extract (GJG) and its representative and bioavailable constituents, geniposidic acid, plantagoguanidinic acid, paeoniflorin, catalpol, loganin, and neoline, on the metabolism of paclitaxel via CYP2C8 and CYP3A4 using pooled human liver microsomes and cultured human cryopreserved hepatocytes to provide the drug information about the pharmacokinetic interaction of this combination therapy. GJG significantly inhibited the production of 3'-p-hydroxypaclitaxel and 6α-hydroxypaclitaxel in vitro in a concentration-dependent manner. The half maximal inhibitory concentration (IC

With the rapid development of electronic health care, the era of medical big data has already emerged. However, in the global electronic health industry environment, one of the significant challenges is that the various medical institutions are independent of one another. Patients, doctors and medical researchers have significant barriers in accessing medical data. As an intervention strategy using blockchain principle, this paper explores the characteristics of blockchain which are applicable to the management of electronic health records (EHRs), and presents a credible blockchain-based electronic health records (CB-EHRs) management platform. A CB-EHRs platform is characterized by decentralization, data tamper-proof, collective maintenance mechanisms, security and credibility. This platform cannot only realize data sharing between medical institutions, but also ensures the privacy of users. This paper introduces the components of the CB-EHRs platformmodel and the implementation principle of its related functions. In addition, this paper also reviews and selects the delegated Byzantine Fault Tolerance (dBFT) consensus mechanism as a viable option for the CB-EHRs platform. Finally, by comparing with the Practical Byzantine Fault Tolerance (PBFT) consensus mechanism and our research, we highlight the potential advantages of our proposed CB-EHRs platform in the medical domain.

This study aims to explore the possibility of using a multilayer artificial neural network for the classification between children with autism spectrum disorder (ASD) and typically developing (TD) children based on short-time spontaneous hemodynamic fluctuations. Spontaneous hemodynamic fluctuations were collected by a functional near-infrared spectroscopy setup from bilateral inferior frontal gyrus and temporal cortex in 25 children with ASD and 22 TD children. To perform feature extraction and classification, a multilayer neural network called CGRNN was used which combined a convolution neural network (CNN) and a gate recurrent unit (GRU), since CGRNN has a strong ability in finding characteristic features and acquiring intrinsic relationship in time series. For the training and predicting, short-time (7 s) time-series raw functional near-infrared spectroscopy (fNIRS) signals were used as the input of the network. To avoid the over-fitting problem and effectively extract useful differentiation features from a sample with a very limited size (e.g., 25 ASDs and 22 TDs), a sliding window approach was utilized in which the initially recorded long-time (e.g., 480 s) time-series was divided into many partially overlapped short-time (7 s) sequences. By using this combined deep-learning network, a high accurate classification between ASD and TD could be achieved even with a single optical channel, e.g., 92.2% accuracy, 85.0% sensitivity, and 99.4% specificity. This result implies that the multilayer neural network CGRNN can identify characteristic features associated with ASD even in a short-time spontaneous hemodynamic fluctuation from a single optical channel, and second, the CGRNN can provide highly accurate prediction in ASD.

Malnutrition and acute kidney injury (AKI) are common complications in hospitalised patients, and both increase mortality; however, the relationship between them is unknown. This is a retrospective propensity score matching study enrolling 46 549 inpatients, aimed to investigate the association between Nutritional Risk Screening 2002 (NRS-2002) and AKI and to assess the ability of NRS-2002 and AKI in predicting prognosis. In total, 37 190 (80 %) and 9359 (20 %) patients had NRS-2002 scores sd 7·8) v. 10·4 (sd 6·2) d, P < 0·05), higher mortality rates (9·6 v. 2·5 %, P < 0·05) and higher incidence of AKI (28 v. 16 %, P < 0·05) than patients with normal nutritional status. The NRS-2002 showed a strong association with AKI, that is, the risk of AKI changed in parallel with the score of the NRS-2002. In short- and long-term survival, patients with a lower NRS-2002 score or who did not have AKI achieved a significantly lower risk of mortality than those with a high NRS-2002 score or AKI. Univariate Cox regression analyses indicated that both the NRS-2002 and AKI were strongly related to long-term survival (AUC 0·79 and 0·71) and that the combination of the two showed better accuracy (AUC 0·80) than the individual variables. In conclusion, malnutrition can increase the risk of AKI and both AKI and malnutrition can worsen the prognosis that the undernourished patients who develop AKI yield far worse prognosis than patients with normal nutritional status.

With the rapid development of computer and communication, data mining and related fields are facing huge challenges. In the research of data mining and text analysis, how to extract concepts more objectively and reasonably is a key issue. Existing approaches seem to lack the intensive analysis of structural nature of structured text, which will affect the effectiveness and accuracy of text analysis. This paper proposes a novel approach of structured text analysis and concept extraction using entropy weight method. It quantitatively assesses the contribution of each module in evaluating the weight of concept. Based on the position of the concept in the text, concepts can be mined more effectively. Furthermore, the method is applied to the field of academic text analysis to verify its effectiveness and advantages over traditional methods. And the experimental results also prove that the method can provide solid theoretical and technical support for personalized academic information services

Cardiac surgery-associated acute kidney injury (CSA-AKI), one of the most severe complications in patients with cardiac surgery, is associated with considerable morbidity, mortality and high costs thus placing a heavy burden to society. Therefore, we aimed to build a predictive model based on preoperative features in order to early recognize and intervene for patients with high risk of CSA-AKI. In this retrospective cohort study, baseline perioperative hospitalization information of patients who underwent cardiac surgery from October 2012 to October 2017 were screened. After multivariate logistic regression, identified independent predictive factors associated with CSA-AKI were incorporated into the nomogram and the discriminative ability and predictive accuracy of the model was assessed by concordance index (C-Index). Additionally, internal validation was performed by using bootstrapping technology with 1000 resamples to reduce the over-fit bias. In all 4395 patients with cardiac surgery October 2012–October 2017, no patients were excluded for the continuous renal replacement therapy (CRRT) before surgery while 2495 patients were excluded due to only one or less than one Scr assay post-surgery. In the end, a total of 1900 patients were enrolled in the study, of which 698 patients (74.89%) developed AKI stage 1, 158 (16.96%) AKI stage 2 and 76 (8.15%) AKI stage 3. After multivariate logistic regression, age, perioperative estimated glomerular filtration rate (eGFR), lactate dehydrogenase (LDH), prothrombin time (PT), with a history of surgery, transfusion, cardiac arrhythmia, coronary heart disease (CHD), or chronic kidney disease (CKD), using calcium channel blocker (CCB), proton pump inhibitors (PPI), non-steroidal anti-inflammatory drugs (NSAID), antibiotic or statin before surgery were predictive factors of CSA-AKI. In addition, the nomogram demonstrated a good accuracy in estimating CSA-AKI, with an C-Index and a bootstrap-corrected one of 0.796 (SD = 0.018, 95% CI 0.795–0.797) and 0.789 (SD = 0.015, 95% CI 0.788–0.790), respectively. Moreover, calibration plots showed an optimal consistency with the actual presence of CSA-AKI. The novel predictive nomogram achieved a good preoperative prediction of CSA-AKI within the Kidney Disease: Improving Global Outcomes (KDIGO) criteria. Though the model, the risk of an individual patient with “subclinical AKI” undergoing cardiac surgery could be determined earlier and such application was helpful for timely intervention in order to improve patient’s prognosis.

Social network media analytics is showing promise for prediction of financial markets. However, the true value of such data is unclear due to a lack of consensus on which instruments can be predicted. In this paper, we investigate whether measurements of collective emotional states derived from large scale network feeds are correlated to the stock transaction data over time. The information space corresponding to stocks is divided into the network public opinion space $$Opinion\_Space$$Opinion_Space and the realistic transaction space $$Behavior\_Space$$Behavior_Space. We then handle the information and generate the multidimensional time series from them respectively. Furthermore, Granger causality analysis and information theory measures are used to find and demonstrate that social media sentiments contain statistically significant ex-ante information on the future prices. At last, we propose our separate-LSTM model and the experimental results of six stocks which are randomly selected indicate that financial data predictions can be significantly improved through our model by the fusion of network public opinion emotions and realistic transaction data.

Due to the difficulties of direct tensile testing on concrete, only limited and often conflicting data are available. In this paper, to investigate the effect of loading frequency on the low cycle fatigue behavior of plain concrete in direct tension, several series of cyclic tension tests at different frequencies and stress levels were carried out. Especially the behavior before the initiation of macroscopic fatigue cracks in concrete, such as irreversible strain accumulation and damage evolution with the number of fatigue cycles through experiments. Test results show that the fatigue life increases with the increasing of the applied loading frequency, while decreases with the increasing of stress level. The stiffness degradation shows a typical three-stage process with the number of fatigue cycles. A new damage model considering the effect of the applied loading frequency and stress level is proposed based on the stiffness degradation and the secondary strain rate during low-cycle fatigue tests. Also, a fatigue model is proposed herein to characterize the deformation and damage behavior of concrete under cyclic tension. The calculated fatigue lives of concrete are compared with experimental data and show good fitting results.

In a split Hopkinson pressure bar bending test on quasi-brittle materials, the specimen maintains a non-equilibrium state and failure occurs before the response of the supports (the transmission bars) appears, as would occur in a support-free impact test. This paper investigates principally the dynamic response of un-notched concrete beams subjected to three-point bending load. To guarantee the accuracy of experimental results analysis, a dynamic response analysis is introduced in the investigation. This assumes that the specimen fractures before the support reaction appears. The long beam model is used in this paper to study the dynamic bending behaviour of specimens that fail after a very short duration of loading. The test results show that the dynamic flexural tensile strength of concrete is strain rate sensitive, and the dynamic increase factor is linearly proportional to the strain rate. Finally, the finite-element software LS-Dyna is used to simulate the failure pattern of specimens in a dynamic bending test. The simulation results verify the validity of the long beam model in processing the experimental data.

In order to explore the dynamic mechanical properties and damage evolution for concrete under the repeated impact loading, the large-diameter (74 mm) Split Hopkinson Pressure Bar (SHPB) was used to conduct the dynamic compression test. The damage was calculated by the Weibull statistical damage model, and the stress-strain curves of concrete were analyzed to study the repeated impact loading effects on the damage of concrete. Meanwhile, the fracture process of the concrete was simulated with the finite element software (LS-DYNA) and Johnson Holmquist Concrete (*MAT_JOHNSON_HOLMQUIST_CONCRETE) constitutive model. The results showed that the dynamic elastic modulus and the peak stress of the specimens decreased with the increased number of repeated impacts, while the peak strain increased. The damage of concrete increased with the increased strain, and the damage increased sharply after the “inflection point”. In addition, the results obtained by finite element simulation agree with the experimental results.

A test method has been developed to obtain reliable post-peak cyclic behavior of concrete in this study. Considering three various loading regimes, strain rate effect on post-peak stress-strain response of concrete is investigated. The envelope curve is found coincides with the monotonic loading curve, which is shown to govern the cyclic response. Concrete presents slightly rate sensitive when it is subjected to static loading. The tensile strength increases with the increasing of strain rate. The stress releases with cyclic even if the maximum strain is constant. It demonstrated that the stress release process is independent of strain rate and it can be expressed by a power function. Analytical expressions introducing damage index are proposed to describe the entire monotonic behavior, as well the response to cyclic loadings. The analytical model uses the concept of uniqueness of envelope curve. The calculated results predicted by the analytical expressions have good agreement with test data.

Due to the difficulties of testing concrete in direct uniaxial tension, only limited data are available. In this paper, the cyclic tensile test on the plain concrete was carried out. Based on the experimental data, the plasticity and damage evolution can be formulated with respect to elastic strain. A constitutive model for cyclic tensile behavior of concrete is proposed. The proposed model can describe the nonlinear post-peak performance of concrete in direct tension including hysteresis loops. Finally, a statistical damage model is derived to describe the microcrack growth of concrete under monotonic and cyclic tension loading. Results demonstrated that the microcracks growth of concrete under cyclic tension loading can be formulated by the Weibull cumulative distribution function.

This paper proposes a method of context-aware edge similarity segmentation algorithm of time series. The algorithm overcomes the machinery of traditional pattern discovery and traditional pattern matching and breaks the limitation of the fixed mode of boundary movement. The algorithm also has stronger adaptive capacity in the humanities sequence activities based on the network. Considering the edge connected by the adjacent points of time series as a fragment, and combining the similarity between the edges calculated by context relation to extract edge density, we can obtain data groups with similar characteristics with the clustering algorithm. This method can not only find the implicit relationships between edges, but also can increase time series segmentation robustness. Experimental results show that our algorithm can reduce the failure rate of segments caused by only considering the pattern matching, and has better anti-interference performance in noisy environments.

Specific network space, including virtual space and practical space, is a space for executing group behavior on specified regions via network. Due to the variability and unpredictability of time series in group behavior in special network space, the detection of normal and abnormal borders faces significant challenges. The parameters in traditional time series mode need to be predefined such as clustering method and anomaly detection methods science the results influentially depend on the selection of parameters. According to the characteristics of data, this paper proposes an efficient method called separation degree algorithm that can construct the self-adaptive interval based on the separation degree model to filter out anomaly data in virtual and practical spaces. The advantage allows us to automatically find the self-adaptive interval to improve the accuracy and applicability of anomaly detection based on the characteristics of the data instead of set parameters of traditional methods in network space. The extensive experimental result shows that the proposed method can effectively detect anomaly data from different spaces.

With the development and popularization of Internet big data, the cyber world nearly parallels to the real world to reflect social hot event. However, the multi-source and heterogeneity features of reality data lead to incompleteness and unreality based on Internet data. To solve the above issues, we propose Event Space Model ESM for event analysis by multi-viewer. In the model, each event is mapped into Network Public Opinion Data Space OS and Actual Behavior Data Space BS. In addition, we propose an Event Starting Point Discovery Algorithm ESPDA to effectively detect starting point of event and to identify the representation of event trajectory in OS and BS. The experimental results prove that the ESM represent event from multi-viewer, while the ESPDA intensively identify events and improve recall and precision of starting point of event. Copyright © 2016 John Wiley & Sons, Ltd.

Objective This paper aims to explore whether a single functional near-infrared spectroscopy (fNIRS) channel can adequately distinguish Autism spectrum disorder (ASD) from typically developing (TD) and independently obtain complete feature information, so as to determine the characteristics of ASD. Method Using the fluctuation entropy of fNIRS from 25 children with ASD and 22 TD children, the pattern of diversity and self-similarity on the time-series was objectively discovered and quantified. By using this characteristic feature and a network cluster classification model consisting of support vector machine (SVM), Gaussian Naive Bayes (GNB), decision tree (DT), multi-layer perceptron (MLP) and logistic regression (LR) to diagnose ASD. Result The fluctuation entropy of TD is greater than that of ASD. The finding, which comes from a mathematical point of view, may reflect the ASD characteristics of higher self-similarity in the spontaneous activity. The classification between ASD and TD was achieved with an accuracy of 97.8% by using the fNIRS time series from one single optical channel. Conclusion The fluctuation entropy of fNIRS could accurately distinguish ASD, even using the time-series from a single optical channel. Significance The fluctuation entropy in this study is a very sensitive index for differentiating between ASD and TD, which might have the penitential to be a fNIRS-based biomarker in the future for the prediction of ASD. In addition, we believe that single channel can provide accurate ASD and TD identification results, which has potential application value in ASD screening in the future.

Background We aimed to develop a nomogram based on preprocedural features for early prediction of acute kidney injury (AKI) and to assess the prognosis in patients after radical and partial nephrectomy. Methods The study included a development cohort of 1111 patients who were treated between June 2012 and June 2017 and an additional validation cohort of 356 patients who were treated between July 2017 and June 2018. Stepwise regression and logistic regression analyses were used to evaluate the association between predictors and AKI. Incorporating all independent predictors, a nomogram for postoperative AKI was developed and externally validated. Patients were followed up for 5 years to assess renal function, acute kidney disease (AKD), chronic kidney disease (CKD), hospital readmission and mortality were key prognosis we focused on. Results After multivariate logistic regression, radical nephrectomy (odds ratio (OR) = 3.57, p

A new calibrated empirical liquefaction triggering correlation equation with five parameters determined by trial and error is derived from the expanded SPT-based database of the Idriss and Boulanger (2010) and Xie (1984) case histories. The new calibrated correlation equation is reasonably conservative for the expanded SPT-based database and is insensitive to reasonable variations in the components of the liquefaction analysis framework adopted in this study. In addition, a probabilistic version of the new calibrated correlation equation expressed in the form of a mapping function that relates the liquefaction probability to a nominal safety factor obtained using the weighted maximum likelihood technique is verified using a weighted empirical probability approach and Monte Carlo simulations. Liquefaction, transition and non-liquefaction zones for clean sands are defined based on different probability contours, i.e., empirical correlations with different liquefaction probabilities. Liquefaction triggering correlations for liquefaction probabilities of 4%, 15% and 25% represent the very low, low, and moderate probability events, respectively, of a liquefiable soil being mistaken for a non-liquefiable soil in the liquefaction potential evaluation.

The rapid growth of web services need efficiently discovering the desired web services for the users. Web service interfaces are defined with WSDL that is described by a bag of terms. Many similarity metrics are proposed to solve this problem, it is hardly to resolve the problem that only few pairs of terms between two services have high semantic distance, the semantic distance of other terms between two services are low. Using traditional keyword search metrics may acquire a wrong result that these two web services are similar, in addition, semantics of the web services is hardly to exploit. In this work the authors firstly help the request service to find the services that belong to the same class, and then they use association rule to find terms that are often appear together and find the most similar terms. The authors weaken the weight of the most similar term contained in an association rule and enhance the other terms' weight contained in an association rule to solve the situation above. The experiments show that our approach outperforms some searching methods.

The purpose of this chapter is to analyze the investor’s psychological characteristics and investment decision-making behavior characteristics, to study the investor sentiment under the network public opinion, and then analyze from three aspects: First, investor sentiment analysis and how to spread in the online media; The influence mechanism of investor’s emotion on the stock market and its effect; the third one is to measure the investor’s emotion based on the degree of attention, trying hard to sort out the internal relations between the investor’s sentiment and the network public opinion and the stock market, in order to lay the theoretical foundation of this article.The purpose of this chapter is to analyze the investor’s psychological characteristics and investment decision-making behavior characteristics, to study the investor sentiment under the network public opinion, and then analyze from three aspects: First, investor sentiment analysis and how to spread in the online media; The influence mechanism of investor’s emotion on the stock market and its effect; the third one is to measure the investor’s emotion based on the degree of attention, trying hard to sort out the internal relations between the investor’s sentiment and the network public opinion and the stock market, in order to lay the theoretical foundation of this article.

Purpose Furosemide is the most widely used diuretic for congestive heart failure and edema; however, the risk of furosemide-associated acute kidney injury (FM-AKI) is unknown. Methods A total of 430,000 inpatients were enrolled in this study. Propensity score matching (PSM) was used to balance covariates across study groups, and then a nomogram was developed to predict FM-AKI. Results The incidence of AKI was 23.94% in the furosemide group and 16.72% in control group after PSM, and the risk of AKI increased with the dose of furosemide. According to the regression analysis, the single day total dose of furosemide was the most important factor for AKI, followed by administration in the intensive care unit (ICU), the estimated glomerular filtration rate, and 13 other indicators. The effect of furosemide in distinct clinical setting is different, subgroup analyses showed that furosemide acted synergistically with surgery and previous treatment with beta-blockers, ACEIs/ARBs and antibiotics to increase the risk of AKI. Then, we developed a prognostic visualized nomogram to predict AKI for furosemide users and found that patients with high scores might have a high risk of AKI; this finding was validated by calibration analyses. The agreement between the nomogram predictions and actual possibility of AKI was excellent, with a probability of 77.40%. Conclusion The use of furosemide in clinical practice requires scrutiny, and the dose, use of other medications and renal function of patients must be taken into account. Our practical prognostic FM-AKI model can help physicians with clinical decision-making and provide treatment guidance. Funding: This work was supported by the National Natural Science Foundation of China (81170688, 81470973, and 81770679). No funding bodies had any role in study design, data collection and analysis, decision to publish, or preparation of the manuscript Declaration of Interest: The authors have declared that no competing interests exist. Ethical Approval: Clinical data were obtained via electronic medical records and a database review and were deidentified to protect patient privacy. The Institutional Review Board approved this study. The need for informed consent was waived for this study.

The study on the interactions between different or the same variables of financial markets is an interesting topic. Many efforts have been devoted to investigate this issue. However, there has been little work studying the relationship of the various attributes within the stock, while this relationship is essential for us to have a deeper understanding of stock’s internal mechanisms. So in this paper, we explored using sequence-to-sequence model for extracting the relationship of arbitrarily two properties of the stock. We not only give a qualitative description of the relationship between stock’s attributes, but also quantify the relationship through the model. The experimental results show that there are certain correlations between the internal attributes of the stock, among which the correlation between \( Close \& \% Tuv\) and \( \% Chg \& \% Tuv\) are more prominent. In addition, we also conducted the anomaly detection on network public opinion information, and found out the starting points of abnormal events combined with the network news information. By comparing the starting points of the events and the changes in the relationship between stock attributes, we concluded that there is a certain regularity between them.

“Digital campus” is an inevitable school pattern of its development in the information age. Under the guidance of the state and localities, some experiences have been gained in the promotion of school informatization with application-driven as the basic principle or working principle. In the present study, digital campus in the first batch of information pilot units of the Ministry of Education was taken as representatives, and their practice was analyzed to understand the nature of application-driven. Meanwhile, this paper further introduced the typical experience of digital campus construction and application driven by three forces—problem-driven, demand-driven and idea-driven. These experiences can provide references for schools in different development stages and different levels of development. Furthermore, the article analyzed issues of the aforementioned different types of power-driven, and proposed suggestions such as highlight the development of “human” ability, give full play to the advantages of “cloud computing”, as well as emphasize network security and localization management and services.

Cardiovascular disease (CVD) caused by atherosclerosis is one of the major causes of death world-wide. Currently, diverse machine learning models have been applied to disease prediction and classification. However, most of them tend to focus on the performance of the algorithm and neglect the underlying variables for patients in different carotid atherosclerotic stages. In this paper, we propose a novel hybrid machine learning method named Subset Learning (S-learning) to predict and discover the risk factors of these different stages. The S-learning algorithm can elucidate the variables that have significant influence on the outcome of carotid atherosclerotic. Performance comparisons are based on the dataset collected from both Shanghai Renji and Shanghai Huashan Hospital. The result shows that the proposed method has superior classification performance than other classification algorithms. Our findings point to the utility of predictive machine learning and the discovery of risk factors to refine the treatment plans.

The formalization of business process with flexibility is to make the collaboration between organizations more detailed and explicit. Based on business process modeling notation, this paper firstly abstracts and concludes explicit and implicit structural features of the across-organizational business process with dynamism and flexibility. Meanwhile, a hierarchical mapping model is then constructed combined with the service interaction patterns. Furthermore, the choreography of transmission on message, channels and parameters between entities is achieved in a modular approach, and defined with the semantic rules of Pi-Calculus. It concludes that the method can effectively distinguish the collaborative relationship between organizations, and also be used to achieve the formal choreography of complicated and dynamic business process.

With the rapid development of information technology, many kinds of personalized information services have come forward. A key issue is how to represent knowledge effectively, which has a great impact on the personalized service quality. Existing approaches seem to lack the cognition characteristics, which makes information services unable to meet the users’ current cognition level. This paper proposes a novel approach of domain knowledge representation based on the specific academic application background, which shows not only the academic concepts in a specific research filed, but also the logic relation between them. It makes the knowledge representation contains abundant semantics. And we propose the concept cognition energy to evaluate the contribution and value of concept to the specific academic domain, which enhances concept’s domain correlation. Furthermore, the approach presents a hierarchical structure according to the concepts’ profession degree in the field, which ensures the knowledge representation to have the characteristic of cognition. Experimental results demonstrate the effectiveness of the method.

This paper proposes a method of the interactive proof to fuse the multi-source text from the internet. The data from the internet described by ontology semantics, and the computing combined the semantic analysis with the intelligent computation. An experiment about the relation of the hypertension and the food made in this paper, which used the former method.

Background Functional near-infrared spectroscopy (fNIRS) was used to investigate spontaneous hemodynamic fluctuations in the bilateral temporal cortices for typically developing (TD) children and children with autism spectrum disorder (ASD). New method This paper proposed an approach to estimate the global time-varying behavior of brain activity through the measurement on change in first-order statistical properties directly from fNIRS time series. Then, a deep learning model combining the long-short term memory (LSTM) and convolutional neural network (CNN) was constructed based on the integration strategy with improved bagging algorithm, with the purpose to explore the potential patterns of temporal variation for ASD identification. Results Based on the theory of stationarity, analysis on the global time-varying behavior of hemodynamic fluctuations in oxy-hemoglobin (HbO2) and deoxy-hemoglobin (Hb) demonstrated that children with ASD showed weaker internal logic, but stronger memory and persistence to random shocks than TD children. Differentiating between ASD and TD with the proposed deep learning approach resulted in high accurate classification with sensitivity of 97.1 % and specificity of 94.3 %. Comparison with existing methods Using fNIRS time series of Hb from single optical channel, we achieved a better classification accuracy of 95.7 % that was about 8 % higher than previous methods with similar data. Conclusions The characterization on time-varying behavior of brain activity holds promise for better understanding the underlying causes to ASD. And the deployed deep learning framework with an integration manner has the potential for screening children with risk of ASD.

Data pre-deployment in the HDFS (Hadoop distributed file systems) is more complicated than that in traditional file systems. There are many key issues need to be addressed, such as determining the target location of the data prefetching, the amount of data to be prefetched, the balance between data prefetching services and normal data accesses. Aiming to solve these problems, we employ the characteristics of digital ocean information service flows and propose a deployment scheme which combines input data prefetching with output data oriented storage strategies. The method achieves the parallelism of data preparation and data processing, thereby massively reducing I/O time cost of digital ocean cloud computing platforms when processing multi-source information synergistic tasks. The experimental results show that the scheme has a higher degree of parallelism than traditional Hadoop mechanisms, shortens the waiting time of a running service node, and significantly reduces data access conflicts.

Interactions between financial time series are complex and changeable in both time and frequency domains. To reveal the evolution characteristics of the time-varying relations between bivariate time series from a multi-resolution perspective, this study introduces an approach combining wavelet analysis and complex networks. In addition, to reduce the influence the phase lag between the time series has on the correlations, we propose dynamic time-warping (DTW) correlation coefficients to reflect the correlation degree between bivariate time series. Unlike previous studies that symbolized the time series only based on the correlation strength, the second-level symbol is set according to the correlation length during the coarse-graining process. This study presents a novel method to analyze bivariate time series and provides more information for investors and decision makers when investing in the stock market. We choose the closing prices of two stocks in China’s market as the sample and explore the evolutionary behavior of correlation modes from different resolutions. Furthermore, we perform experiments to discover the critical correlation modes between the bull market and the bear market on the high-resolution scale, the clustering effect during the financial crisis on the middle-resolution scale, and the potential pseudo period on the low-resolution scale. The experimental results exactly match reality, which provides powerful evidence to prove that our method is effective in financial time series analysis.

The need for high-resolution imaging becomes particularly important in remote sensing image applications, such as ground-object identification. We introduce a Bayesian multiframe super-resolution algorithm that efficiently improves the imaging resolution of our micro–nano carbon satellite images. We begin by presenting a theoretical overview of the algorithm, and subsequently we conduct experiments in two phases. In the first phase, we compare the performance of our algorithm with those of other similar algorithms using a set of reference images. In the second practical application phase, we apply all the algorithms considered to panchromatic images obtained from our civilian micro–nano carbon satellite. Our results indicate that the proposed algorithm significantly outperforms the other algorithms, affording higher image resolution and greater image detail. We believe that our approach can contribute to the further development of satellite imaging systems.

Based on the features of marine environmental data and processing requirements, a cloud computing architecture of marine environment information is proposed, which provides a new cloud technology framework for the integration and sharing of marine environmental information resources. A physical layer, software platform layer and an application layer are illustrated systematically, at the same time, a corresponding solutions for many difficult technical problems such as parallel query processing of multi-dimensional, spatiotemporal information, data slice storage, software service flow customization, analysis, reorganization and so on. A prototype system is developed and many different data-size experiments and a comparative analysis are done based on it. The experiment results show that the cloud platform based on this framework can achieve high performance and scalability when dealing with large-scale marine data.

Time series classification is an important task in time series analysis. Thus, many methods have been developed for the task. However, the quality of features is difficult to measure and there is no distance measurement method for most areas. And these methods cannot extract the long-term dependency feature from time series. In order to solve these problems, we propose a new time series classification model, Long short-term memory networks and Convolution Neural Networks (LCNN). First, the model can automatically extract features from the time series. Second, LCNN solves the long-term dependence problem by introducing Long short-term memory networks (LSTM) into time series classification tasks. Third, LCNN adopts multi-branch structure to down-sampling and Gaussian noise to process the original time series, which improves the classification performance. In addition, we use the Hurst exponent to measure the long-term dependency in time series. All experiments show that LCNN improves the classification performance and is well suited for small datasets.

Among various index properties used by civil engineers, the uniaxial compressive strength (UCS) of concrete cores is probably the most widely used property of concrete. The quasistatic standard Brazilian disk test and split Hopkinson pressure bar (SHPB) dynamic compression test of the concrete cores are investigated in this paper, and the influence of different initial static loadings on the dynamic compression performance of concrete is investigated. The initial static loadings are divided into five grades, 0, 25, 50, 75, and 90% of the static splitting strength of concrete cores, respectively. Meanwhile, the Weibull statistical distribution model is introduced to analyze the test results of strength, which show that the dynamic compressive strength of concrete increases with the increase of the initial static loading and strain rate. In addition, according to the results obtained from the experimental data, a predication model for dynamic compressive strength of concrete cores is proposed.