An Application of Deep Belief Networks in Early Warning for Cerebrovascular Disease Risk

To reduce the incidence of cerebrovascular disease and mortality, identifying the risks of cerebrovascular disease in advance and taking certain preventive measures are significant. This article was aimed to investigate the risk factors of cerebrovascular disease (CVD) in the primary prevention, and to build an early warning model based on the existing technology. The authors use the information entropy algorithm of rough set theory to establish the index system suitable for early warning model. Then, using the limited Boltzmann machine and direction propagation algorithm, the depth trust network is established by building and stacking RBM, and the back propagation is used to fine-tune the parameters of the network at the top layer. Compared with the LM-BP early-warning model, the deep confidence network model is more effective than traditional artificial neural network, which can help to identify the risk of cerebrovascular disease in advance and promote the primary prevention.

Assessment of early warning signs in hospital projects' front-end phase

PurposeThe front-end phase plays an important role in achieving project success, and establishment of performance measurement systems considering project challenges or pitfalls is a way of keeping track of this phase. Early warning signs, a type of proactive performance indicators, may serve as means for improving decision-making and project processes aiming for short- and long-term project success. In this paper, the authors present findings from a study on early warning signs (EWS) in hospital projects' front-end. A preliminary systematisation of identified signs as a contribution to front-end improvement is provided.Design/methodology/approachThe paper is based on a mixed methods approach, using a sequential, exploratory research design comprising document studies, interviews and a survey.FindingsThe authors identified 62 challenges for hospital projects' front-end performance and further established four categories of EWS as follows: (1) structure and tools, (2) context and frame factors, (3) management and (4) relational factors and properties. This mirrors the presence of hard and soft issues from previous studies. There is need for clarifying terminology and raising consciousness on EWS. Processual approaches to identify EWS are considered more useful than subsequent established indicators.Originality/valueThe findings from this paper provide insight into EWS in hospital projects' front-end phase. This adds to the general understanding of EWS and contributes to more knowledge on the front-end phase in general.

Sustainable development early warning and financing risk management of resource-based industrial clusters using optimization algorithms

PurposeThe purpose is to analyze and discuss the sustainable development (SD) and financing risk assessment (FRA) of resource-based industrial clusters under the Internet of Things (IoT) economy and promote the application of Machine Learning methods and intelligent optimization algorithms in FRA.Design/methodology/approachThis study used the Support Vector Machine (SVM) algorithm that is analyzed together with the Genetic Algorithm (GA) and Ant Colony Optimization (ACO) algorithm. First, Yulin City in Shaanxi Province is selected for case analysis. Then, resource-based industrial clusters are studied, and an SD early-warning model is implemented. Then, the financing Risk Assessment Index System is established from the perspective of construction-operation-transfer. Finally, the risk assessment results of Support Vector Regression (SVR) and ACO-based SVR (ACO-SVR) are analyzed.FindingsThe results show that the overall sustainability of resource-based industrial clusters and IoT industrial clusters is good in the Yulin City of Shaanxi Province, and the early warning model of GA-based SVR (GA-SVR) has been achieved good results. Yulin City shows an excellent SD momentum in the resource-based industrial cluster, but there are still some risks. Therefore, it is necessary to promote the industrial structure of SD and improve the stability of the resource-based industrial cluster for Yulin City.Originality/valueThe results can provide a direction for the research on the early warning and evaluation of the SD-oriented resource-based industrial clusters and the IoT industrial clusters, promoting the application of SVM technology in the engineering field.

Challenges Associated with Creeping Disasters in Disaster Risk Science and Practice: Considering Disaster Onset Dynamics

In this article, we set out to reconcile a general conceptualization of disaster temporalities by drawing on the epitome example of a creeping disaster, namely famine. Our argument is driven by the recognition that slowly manifesting disaster impacts pose distinct challenges for decision makers and researchers while there is a tendency for the disaster literature to overlook the role of disaster onset dynamics. More specifically and as a starting point, we identify four key themes that merit particular attention when dealing with creeping disasters: (1) our understanding of disaster as a phenomenon; (2) measurement and operationalization; (3) early warning and response; and (4) disaster management and termination. By integrating conceptual discussions of disaster with famine scholarship—a phenomenon often excluded from mainstream disaster research—this article provides fresh perspectives on disaster science as well as a number of implications for how we think about disaster risk reduction.

Multi-Source Information Monitoring Test of Fractured Rock Mass Destruction Characteristics and Sensitivity Analysis of Precursor Phenomena

The accuracy of the monitoring information is particularly important for exploring fractured rock mass deformation and failure mechanisms and precursor characteristics. Appropriate monitoring methods can not only timely and effectively reflect the failure laws of fractured rock masses but also play an early warning role. To explore more reasonable monitoring methods, uniaxial compression experiments and real-time non-destructive monitoring on prefabricated fractured rock specimens through DIC, AE, and IRT were conducted; the strain field, temperature field, ringing frequency, standard deviation, etc. were analyzed; and correlation between the three methods in the information of audience was explored. The results show the following. (1) The failure evolution process of fractured rock mass can be divided into four stages. DIC can detect the initiation and propagation of cracks near the fractures of the specimen at the earliest stages. (2) The order of occurrence of precursor phenomena in multi-source monitoring information is different, which is vertical strain field > shear strain field > horizontal strain field > temperature field > ringing times. (3) The dispersion degree of standard deviation of each field is obviously different; the infrared temperature field is greater, but the strain field and temperature field show the same trend. (4) There are obvious precursors before the specimen is on the verge of instability; acoustic emission detected two consecutive increases in the cumulative number of ringing before destruction, which means the most obvious precursors. The research results can provide a theoretical basis for the precursory information capture and damage early warning of the fractured rock mass destruction process.

Nutritional Changes and Early Warning of Moldy Rice under Different Relative Humidity and Storage Temperature

Processed unhusked rice is prone to mildew during storage. In this study, the storage conditions were simulated at temperatures of 20, 30, and 35 °C and a relative humidity of 40%, 60% and 80%, respectively. The water, fatty acid, and total starch content and the peak viscosity, mold colony number, protein secondary structure, and spatial structure of rice were monitored in order to propose the critical point of mildew during storage. In the process of rice from lively to moldy, the water content, fatty acid contents and the peak viscosity were increased. The total starch content decreased and then showed a slow increasing trend, while the microstructure of the powder particles changed from smooth and complete to loosen and hollow. With the increase in storage time, the vibration of the amide Ⅰ band of the rice samples decreased slightly, indicating that the total contents of β-fold, β-turn, α-helix, and random curl of the rice protein also changed. PCA (Principal Component Analysis) analysis showed that rice mildew index was closely related to temperature and humidity during storage. In our investigation, the best and most suitable temperature and relative humidity for rice storge is 20 °C and 40%, respectively. These results suggested that temperature and environmental humidity are vital factors affecting the physicochemical properties and nutrient changes, which provides a theoretical basis for the early warning of rice mildew during storage.