Real-time Predictive Analytics Research Articles

Building Flexible, Data-driven Framework for Real-time Analysis

In the contemporary business landscape, the escalating demand for real-time predictive analytics is driven by the imperative for dynamic decision-making. Traditional analytics models often prove inadequate in addressing the need for agility required to respond swiftly to rapidly changing circumstances. Real-time predictive analytics, however, offers a transformative solution, empowering organizations to make informed and timely decisions in fast-paced environments. This capability proves invaluable in industries were staying ahead of emerging trends is critical, fostering a proactive approach to decision-making that can significantly impact competitiveness. The sheer volume and diversity of data require sophisticated solutions for processing and analysis. Real-time predictive analytics becomes an indispensable tool, offering the capability to promptly extract valuable insights from massive datasets. This not only enhances decision-making but also allows organizations to stay ahead by uncovering trends and patterns in real time. Scalability is a fundamental consideration for organizations on a growth trajectory. Real-time predictive analytics frameworks provide a scalable foundation, allowing businesses to seamlessly expand their analytical capabilities. This adaptability ensures that the framework can handle the increasing demands for processing power and storage, aligning with the evolving needs of a growing organization.

Explainable real-time predictive analytics on employee workload in digital railway control rooms

Both workload peaks and lows contribute to lower employee well-being. Predictive employee workload analytics can empower management to undertake proactive prevention. For this purpose, we develop a real-time machine learning framework to predict and explain future workload in a challenging environment with variable and imbalanced workload: the digital control rooms for railway traffic management of Infrabel, Belgium’s railway infrastructure company. The proposed two-stage methodology leverages granular data of workload categories that are very different in nature and separates the effects of workload presence and magnitude. In this way, the set-up addresses the changing workload mix over 15-minute intervals. We extensively benchmark machine learning and deep learning models within this context, leading to LightGBM (Light Gradient Boosting Machine) as the best-performing model. SHAP (SHapley Additive exPlanations) values highlight the benefits of disentangling presence and magnitude and reveal associations with human-machine interaction and team exposure. As a proof of concept, our implemented predictive model offers tailored decision support to the traffic supervisor in an explainable way. In particular, the tool depicts overloaded and/or underloaded workstations and provides in-depth insights through local SHAP values.

A digital twin in transportation: Real-time synergy of traffic data streams and simulation for virtualizing motorway dynamics

The introduction of digital twins is expected to fundamentally change the technology in transportation systems, as they appear to be a compelling concept for monitoring the entire life cycle of the transport system. The advent of widespread information technology, particularly the availability of real-time traffic data, provides the foundation for supplementing predominated (offline) microscopic simulation approaches with actual data to create a detailed real-time digital representation of the physical traffic. However, the use of actual traffic data in real-time motorway analysis has not yet been explored. The reason is that there are no supporting models and the applicability of real-time data in the context of microscopic simulations has yet to be recognized. Thus, this article focuses on microscopic motorway simulation with real-time data integration during system run-time. As a result, we propose a novel paradigm in motorway traffic modeling and demonstrate it using the continuously synchronized digital twin model of the Geneva motorway (DT-GM). We analyze the application of the microscopic simulator SUMO in modeling and simulating on-the-fly synchronized digital replicas of real traffic by leveraging fine-grained actual traffic data streams from motorway traffic counters as input to DT-GM. Thus, the detailed methodological process of developing DT-GM is presented, highlighting the calibration features of SUMO that enable (dynamic) continuous calibration of running simulation scenarios. By doing so, the actual traffic data are directly fused into the running DT-GM every minute so that DT-GM is continuously calibrated as the physical equivalent changes. Accordingly, DT-GM raises a technology dimension in motorway traffic simulation to the next level by enabling simulation-based control optimization during system run-time that was previously unattainable. It, thus, forms the foundation for further evolution of real-time predictive analytics as support for safety–critical decisions in traffic management. Simulation results provide a solid basis for the future real-time analysis of an extended Swiss motorway network.

The impact of the resumption of in-person school attendance on COVID-affected child abuse and neglect trends in Florida

BackgroundThe COVID-19 pandemic has led to fewer child abuse and neglect (CAN) hotline calls, CAN investigations, and foster care entries across the U.S. ObjectiveTo determine if there were decreases in CAN hotline calls, CAN investigations, foster care entries, and foster care exits in Florida (the largest among the few states that publishes monthly public data on the four areas) after the pandemic began, and to determine if there was any amelioration of these trends in Florida once schools had reopened for in-person learning. MethodsSecondary data analyses of administrative child welfare data from January 2010 to June 2021 from Florida were used. Spline regression equations were calculated for CAN hotline calls, CAN investigations, foster care entries, and foster care exits during three periods: 1) before the COVID-19 pandemic (January 2010 to February 2020); 2) after the pandemic, but before schools reopened (March 2020 to October 2020); and 3) after the pandemic and after schools reopened (November 2020 to June 2021). ResultsPrior to the pandemic, there was a non-statistically significant increase of 1.35 hotline calls per 100,000 children per month (p = .478), a statistically significant increase of 0.01 investigations per 100,000 children per month (p < .001), a non-statistically significant increase of 0.01 foster care entries per 100,000 children per month (p = .415), and a statistically significant increase of 0.03 foster care exits per 100,000 children per month (p < .05). Once the pandemic started in March 2020, there were statistically significant decreases of 136.02 hotline calls per 100,000 children per month (p < .001), 102.84 investigations per 100,000 children per month (p < .001), 6.32 foster care entries per 100,000 children per month (p < .001) and 5.75 foster care exits per 100,000 children per month (p < .01). Once all schools reopened for in-person learning in November 2020, there continued to be statistically significant decreases of 47.86 hotline calls per 100,000 children per month (p < .05), 6.38 foster care entries per 100,000 children per month (p < .001) and 6.53 foster care exits per 100,000 children per month (p < .001). This suggests that there were an estimated 34,374 fewer CAN hotline calls, 2338 children who did not enter foster care, and 2587 youth residing in foster care (YRFC) whose foster care exits were delayed. The delay in foster care exits suggests that YRFC in Florida had stayed a combined cumulative equivalent of 477.1 years longer in care. ConclusionWith the COVID-19 variants like Omicron continuing to wreak havoc in Florida, there will be a continued trend of decreasing CAN hotline calls, foster care entries and foster care exits even with in-person learning in all Florida schools. A real-time interoperable data system utilizing real-time predictive analytics must be developed in concert with the development of leaders and executives with advanced degrees in child welfare organizations who are able to maximize information from such systems.

Deep Learning-based Ethical Judgments in Connected Vehicle Technologies: Route Planning Algorithms, Spatial Data Visualization Tools, and Real-Time Predictive Analytics

Deep Learning-based Ethical Judgments in Connected Vehicle Technologies: Route Planning Algorithms, Spatial Data Visualization Tools, and Real-Time Predictive Analytics

Riding the Digital Product Life Cycle Waves towards a Circular Economy

Data driven organizations such as Amazon and Uber have raised the capabilities and expectations of customers to a new level by providing faster and cheaper products and services. The reviewed literature documented that 10–15% of the online products are returned and in many cases such products are not shelf-ready due to product obsolescence or slight wear and tear, thereby reducing profits. Many of these products are disposed of in landfills. There were very few publications that documented how integration of digitized product life cycle into the business model improves product returns and the remanufacturing processes. As societies continue on, environmentally responsible, digital journeys with connected devices and people, reverse supply chains and remanufacturing will play increased importance in fulfilling customers expanded expectations. The networks are evolving, wherein, data are collected from all phases of the product lifecycles from design, prototype, manufacturing, usage aftermarket, returns remanufacturing and recycling. The objective of this paper’s authors was to describe how all phases of product life cycles can be digitized to improve global reverse supply chains and remanufacturing. The authors performed a literature review and developed case studies to document current and to predict future transformational waves that will become increasingly used in many industrial sectors. The authors made recommendations about the importance of improved product design, reduced processing costs and increased use of remanufactured products based upon data on returns to manufacturers and service providers. This paper contributes to research by providing a framework of a digitized product life cycle integrated with the business process phases including remanufacturing and supported with real-world case studies for practitioners and academicians. The authors outlined potential future topics for academic researchers and practitioners, for expanding usage of digital tools in real-time predictive analytics to improve remanufacturing system’s efficiency and quality.

On Construction of Sensors, Edge, and Cloud (iSEC) Framework for Smart System Integration and Applications

Intelligent systems influence many aspects of daily life. With the emergence of the Internet of Things (IoT), artificial intelligence (AI), and machine learning (ML), opportunities have been created for smart computing infrastructure. However, problems might arise from the lack of interconnectivity, higher reliability, real-time predictive analytics, and low-latency requirements. Therefore, in this article, we propose the sensors, edge, and cloud (iSEC) framework. The project deploys a smart cloud edge-computing architecture to provide ML and deep learning in the cloud edge environment. Two pilot projects of air quality monitoring system and object detection are demonstrated to evaluate the iSEC framework.

Robolution: Real Time Predictive Analytics for Industrial Robots

Robots have been playing a very important role in our day-to-day lives and will be a necessity in the coming future. Whenever we hear automation, the first thing that strikes our mind is a robot performing the given task. But if a robot fails to do the task, it could cost an individual or corporate a huge financial loss. In this study, we have learned the working of various robots and drawbacks that hold them back. For this work, we did make a study of drives used in the robot and after that applied the machine learning algorithms to predict the classification of whether the robot will function properly or not, based on the data of drive(s).

Comparing clinical judgment with the MySurgeryRisk algorithm for preoperative risk assessment: A pilot usability study

BackgroundMajor postoperative complications are associated with increased cost and mortality. The complexity of electronic health records overwhelms physicians’ abilities to use the information for optimal and timely preoperative risk assessment. We hypothesized that data-driven, predictive-risk algorithms implemented in an intelligent decision-support platform simplify and augment physicians’ risk assessments. MethodsThis prospective, nonrandomized pilot study of 20 physicians at a quaternary academic medical center compared the usability and accuracy of preoperative risk assessment between physicians and MySurgeryRisk, a validated, machine-learning algorithm, using a simulated workflow for the real-time, intelligent decision-support platform. We used area under the receiver operating characteristic curve to compare the accuracy of physicians’ risk assessment for six postoperative complications before and after interaction with the algorithm for 150 clinical cases. ResultsThe area under the receiver operating characteristic curve of the MySurgeryRisk algorithm ranged between 0.73 and 0.85 and was significantly better than physicians' initial risk assessments (area under the receiver operating characteristic curve between 0.47 and 0.69) for all postoperative complications except cardiovascular. After interaction with the algorithm, the physicians significantly improved their risk assessment for acute kidney injury and for an intensive care unit admission greater than 48 hours, resulting in a net improvement of reclassification of 12% and 16%, respectively. Physicians rated the algorithm as easy to use and useful. ConclusionImplementation of a validated, MySurgeryRisk computational algorithm for real-time predictive analytics with data derived from the electronic health records to augment physicians’ decision-making is feasible and accepted by physicians. Early involvement of physicians as key stakeholders in both design and implementation of this technology will be crucial for its future success.

Elaborating the Human Aspect of the NIST Framework for Cyber-Physical Systems.

The National Institute of Standards and Technology (NIST) has developed a Framework for Cyber-Physical Systems (CPS Framework) that supports system engineering analysis, design, development, operation, validation and assurance of CPS. Cyber-physical systems (CPS) comprise interacting digital, analog, physical, and human components engineered for function through integrated physics and logic. For instance, a city implementing an advanced traffic management system including real-time predictive analytics and adaptation/optimization must consider all aspects of such a CPS system of systems' functioning and integrations with other systems, including interactions with humans. One Aspect (or grouping of stakeholder concerns) of the CPS Framework is the Human Aspect. NIST is engaging HFES in a panel discussion to elaborate Human Aspect concerns, especially relevant constructs, measures, methods, and tools.

Multi-criteria evaluation of real-time key performance indicators of supply chain with consideration of big data architecture

One of the major issues a designer of Big Data Architecture has to trade with is incorporating real-time predictive analytics capability using offline synergistic approaches like simulation, fuzzy analytic network process, and Technique for Order Preference. Further, under this setting, which involves re-engineering of operational units, the present study proposes a simple, yet practical heuristic to quickly handle the unstructured relational key-performance-indicators (KPIs) data of a supply chain that are obtained from the results of the simulation. Within the big data framework, the proposed model can be used as a decision support tool by the companies to evaluate their KPIs in a real-time dynamic system.

Kafka, Samza and the Unix philosophy of distributed data

Apache Kafka is a scalable message broker, and Apache Samza is a stream processing framework built upon Kafka. They are widely used as infrastructure for implementing personalized online services and real-time predictive analytics. Besides providing high throughput and low latency, Kafka and Samza are designed with operational robustness and long-term maintenance of applications in mind. In this paper we explain the reasoning behind the design of Kafka and Samza, which allow complex applications to be built by composing a small number of simple primitives – replicated logs and stream operators. We draw parallels between the design of Kafka and Samza, batch processing pipelines, database architecture, and the design philosophy of Unix.

Real-Time Predictive Analytics, Big Data &amp;amp; Energy Market Efficiency: Key to Efficient Markets and Lower Prices for Consumers

The combination of evolving deregulation of the US and EU energy markets together with recent advances in data analytics and so called 'Big Data' technologies now offers an unprecedented opportunity for optimal real-time energy pricing for buyers and sellers alike. The main challenge to date for optimal pricing has been optimal real-time bidding and variety of traditional data analysis tools have been applied to this challenge. Yet inefficiencies remain due to the volatile nature of the real-time market. Energy data science is the best solution to protect consumers against the electricity market's inefficiencies. This field is the meeting point between computer programming, machine learning, big data, quantitative analysis and economics. Energy data science is used to help the consumer predict what price they should offer to buy at each hour and closes gaps in the electric markets.