01
Machine learning techniques are used to process highly unbalanced data to classify fatality rates in car accidents
Handling highly imbalanced data for classifying fatality of auto collisions using machine learning techniques
Shengkun Xie & Jin Zhang
Summary
Accurate prediction of fatal events in car accidents is of great significance for health management. This study explores how unbalanced data processing techniques can be applied in machine learning to improve prediction performance. By applying these technologies to crash data, health organizations can identify and predict fatal events, allowing them to allocate limited medical resources more efficiently. At the same time, improving the performance of machine learning models through unbalanced data processing techniques can affect health management decisions. Our findings highlight the importance of imbalance data processing techniques in predicting the lethality of car crashes, ultimately contributing to improved road safety and better management of health resources. In addition, the effective use of imbalanced data can significantly improve the specificity of predictions. Addressing the impact of machine learning techniques on unbalanced crash data can significantly improve overall health outcomes.
Scan the QR code to view the full article
02
Review: Big data analysis models for evaluating non-pharmacological interventions for the pandemic
A review of big data analytics models for assessing non-pharmaceutical interventions for COVID-19 pandemic management
Fatemeh Nawazi, Yufei Yuan & Norm Archer
Summary
During the pandemic, non-pharmacological interventions (NPIs) were the only solution to mitigate COVID infections until vaccines were available. Even after vaccinations began, governments continued to use NPIs. In this study, we reviewed various big data analytics models used to evaluate and optimize the effectiveness of NPI. These models fall into three categories: descriptive analysis, which measures changes in NPI-induced infection rates; predictive analytics, which predicts the future of the pandemic through the implementation of several NPIs; Data-driven, prescriptive analysis to propose the best control policies. We further analyze the underlying assumptions, limitations, and applicability of each approach in different pandemic phases and under different scenarios. This review of the NPI assessment methodology will help policymakers understand which model to choose for policymaking during the pandemic. Finally, we propose some future research directions.
Scan the QR code to view the full article
03
Language hesitant for enterprise resource planning (ERP) selection is fuzzy interactive variant group decision-making
Linguistic hesitant fuzzy interactive multi-attribute group decision making for enterprise resource planning selection
Shuping Wan, Chunyan Zeng, Jiuying Dong & Sishi Hu
Summary
Enterprise resource planning (ERP) system selection involves multiple evaluation attributes that interact with each other. It boils down to an interactive multi-attribute group decision (MAGDM). The Linguistic Hesitant Fuzzy Set (LHFS) is a powerful tool for expressing the uncertainty, hesitation, and inconsistency of decision-makers' preferences. In this paper, two new LHFS interactive MAGDM methods based on integrated exponential geometric aggregation operators are proposed. First, the synthetic cloud of LHFS is defined and a distance metric between the two synthetic clouds is provided. Considering the interaction between aggregate LHFS, we have developed some comprehensive exponential geometric aggregation operators for LHFS. First, we define the uncertainty of LHFS. Then, we propose a method to derive the weights of decision-makers and a method to derive the weights of composite attributes. Therefore, this paper proposes two new LHFS interactive MAGDM methods. Finally, this paper provides an example of ERP selection to validate the proposed approach.
Scan the QR code to view the full article
04
Emergency production of medical products: partial and full decentralization
Emergency production of medical products: partial decentralization vs. complete decentralization
Huaige Zhang, Yimeng He, Yuke Hu, Yeming Gong, Xianpei Hong & Li Zhang
Summary
In the emergency production of medical products, the supplier with the raw materials is often the leader in the game and has the advantage of determining the structure of the supply chain. Previous research on supply chain structure has focused on how leaders can leverage this advantage for optimal profits, while ignoring followers' strategic compensation for this disadvantage. This article explores how followers can make the best profits in a supply chain consisting of two buyers (integrated and non-integrated buyers) and one supplier through contract design and production timing. In the event of a shortage of medical products, enterprises form a fully decentralized or partially decentralized (supplier and buyer form a vertically integrated entity) structure for emergency production. The results show that in a partially decentralized structure, the integrated buyer can make more profits. In addition, in a partially decentralized structure, non-integrated buyers are more sensitive to production time. Surprisingly, in emergency production, the more products a non-consolidated buyer buys, the higher the supplier's wholesale price. However, under certain conditions, this anomaly increases the likelihood that the buyer will make the best profit and weakens the impact of the supply chain structure.
Scan the QR code to view the full article
05
Development of a start-to-end matrix-based business analysis for BRT systems using smart card data: the case of Jakarta, Indonesia
Using smart card data to develop origin-destination matrix-based business analytics for bus rapid transit systems: case study of Jakarta, Indonesia
Meditya Wasesa, Mochammed Agus Afrianto, Fakhri Ihsan Ramadhan, Yos Sunitioso, Shimaditya Nooraini, Uttam Surgeon Son & Sri Hastuti
Summary
Bus Rapid Transit (BRT) systems have always been an indispensable mainstay of public transport, especially in densely populated areas. In the context of BRT resource allocation planning, it is important to have an accurate understanding of the utilization of the BRT network. The research focuses on how operators can develop business analytics based on the Origin and Destination Point (OD) matrix using passengers' smart card data. This study proposes a hybrid method that combines trip linking, direct pairing, pattern estimation methods, and visual analysis development. This novel approach is robust in handling incomplete smart card data transactions, generating OD matrices and corresponding visual analytics that serve as a decision support system for BRT operators. We applied the proposed analysis method to more than 20.6 million smart card transaction data of one of the largest global BRT systems in Jakarta, Indonesia, and validated it.
Scan the QR code to view the full article
06
Analysis of ESG information disclosure strategies for investors with different risk appetites
Analysis of ESG information-disclosure strategies considering different risk-appetite investors
Peilun Li, Jiali Zhu & Ye Zhou
Summary
In response to the investment market and the public, as of 2022, more than 1,000 listed companies in China have started voluntarily providing environmental, social and governance (ESG) information. How to establish an effective ESG information disclosure mechanism is important for reliable external financing providers. Here, we design a risk management model to describe the information disclosure strategy decisions of providers under different risk appetite investors, and analyze how information effectiveness, investors' information sensitivity and risk aversion, and rating agency accuracy affect the optimal ESG information disclosure strategy. We find that: (1) providers can reduce their information collection and disclosure costs by adopting appropriate inadequate disclosure strategies, thereby maximizing the benefits of ESG information disclosure; (2) investors' information sensitivity and risk aversion will affect the space for enterprises to adopt insufficient disclosure strategies; (3) More professional rating agencies can provide stronger credit endorsement for ESG information, prompting providers to reduce the disclosure of information.
Scan the QR code to view the full article
07
TOPSIS is a three-way decision on food business site selection under the benefit-opportunity-cost-risk platform
Three-way decision TOPSIS for food business location under a benefit-opportunity-cost-risk platform
Shone Andre Scott, Selbert Heimang, Rosen Ancheta, JR, Fatima Maturan, Joerabel Lourdes Aro, Samantha Shane Evangelista, Nadine May Attebing & Landon Ocampo
Summary
This study comprehensively assesses the factors that can help policymakers find the best position for food businesses in developing economies. It incorporates analytic hierarchy process (AHP) in determining standard weights (i.e., benefits, opportunities, costs, risks) and adopts the newly introduced three-way decision-ideal solution similarity ranking method (3WD-TOPSIS) in determining priority factors. The application of AHP allows the priority to be allocated more to benefits and opportunities than to costs and risks, reflecting the benefit- or opportunity-oriented attitude of food companies. At the same time, through the implementation of 3WD-TOPSIS, it was found that government regulations and restrictions, proximity to consumers, parking capacity, supply chain strategy, and socioeconomic status were the most critical location decision-making factors. The results of the comparative analysis showed that the results of this study were highly consistent with those of other comparable methods. Finally, this paper outlines the management implications of the findings.
Scan the QR code to view the full article
08
Rating Disclosure Analysis: The Impact of Compensation Reform on Reducing Credit Rating Bias
Rating disclosure analysis: compensation reform consequence on counteracting biased credit ratings
Kittiphod Charuntham, Kesar Kanchanpum, Nartafi Tanko & Jirawat Vorkantak
Summary
This study analyzes a proposed compensation reform aimed at preventing credit rating agencies from generating inaccurate ratings. Specifically, this paper explores the motivations for credit rating agencies to observe portfolio signals and adopt rating disclosure policies under rating contracts, based on rating contracts and incentive contracts, respectively. The issuer has a risk portfolio and solicits ratings from credit rating agencies, which endogenously observe signals and decide on rating disclosure policies during the rating process. The results of the study suggest that under rating-based contracts, credit rating agencies do not make an effort to observe signals and exaggerate ratings. Under incentive contracts, credit rating agencies always make an optimal effort to observe signals and adopt a full disclosure regime. As a result, incentive contracts are more incentivizing credit rating agencies to make more efforts to improve rating accuracy and implement a full disclosure policy than rating-based contracts.
Scan the QR code to view the full article