A hybrid machine learning-enhanced MCDM model for transport safety engineering

A hybrid machine learning-enhanced MCDM model for transport safety engineering

Dynamic evolution of SPIs

To analyze the evolution of specific SPIs over time, we created radar plots, as depicted in Fig. 8. In the radar plots, each axis corresponds to a different country, with the distance from the center reflecting that country’s performance score changes on a specific criterion. Tracking these changes enables ongoing performance assessment and adaptation to evaluate conditions. For benchmarking, decision-making, or policy evaluation, recognizing dynamic shifts is essential to respond to trends, enhance performance, and adjust strategies effectively.

Fig. 8
figure 8

Dynamic evolution of the SPIs over the past ten years.

Figure 8 illustrates that most countries within the OAS consistently demonstrated improvements in transport safety performance across various indicators. This trend underscores the increasing focus on transport safety development among OAS member countries. For example, Guyana (GY) displayed significant improvement in indicator A2, whereas El Salvador (SV) and Saint Lucia (LC) exhibited notable improvements in indicator A3. Most countries exhibit continuous improvements in indicators D1, E1, and E2. For indicator D2, which measures GDP per capita, all nations except Canada (CA) and the United States (US) remained relatively unchanged over the past decade. However, certain countries experienced notable regressions in specific areas. The Dominican Republic (DO) showed a decline in indicator A1, and Haiti (HT) regressed on indicator A2. Additionally, for indicators B2 and B3, which measure the percentage of seatbelt usage in the front and rear seats, respectively, most states displayed a regressive trend. The most significant declines were observed in Peru (PE) for B2 and Saint Vincent and the Grenadines (VC) for B3.

Deconstruction of composite score

To understand the impact of each SPI on the overall index, we divided the index into individual components, as depicted in Fig. 9. This division aids in analyzing the strengths and weaknesses of the index, identifying key drivers of the index score, and assessing the relevance of each component.

Fig. 9
figure 9

Breaking down the overall safety score to highlight the specific contributions of individual indicators.

As depicted in Fig. 9, AG, BB, and VC achieved the highest overall safety scores. Several key indicators significantly contributed to these high scores. Notably, AG was uniquely influenced by indicator B1, which measures the percentage of traffic deaths involving alcohol. Other critical indicators included A2, assessing traffic fatalities; B2, B3, and B4, evaluating the percentage of seatbelt usage in front and rear seats and helmet usage, respectively; and C2, which gauges road network density. Additional influential indicators were D2, reflecting the socioeconomic levels of the nations and underpinning funding for transport safety strategies, and E1E4, which assess the enforcement of various traffic laws, thereby enhancing the transport safety performance in these three countries.

In addition to the top three countries, several nations demonstrated commendable performance on specific indicators that did not significantly impact the leading performers. For instance, KN, ranked fourth, excelled in indicator A1. BR, ranked twelfth, was notable for A3. These findings highlight the necessity for enhanced transport safety strategies and their implementation, even in high-performing countries. This situation offers opportunities for shared learning and improvement across nations. Moreover, these indicators emphasize the urgent need for countries to strengthen transport safety measures, reduce fatality rates, and expand road network density to ensure safer roads. Overall, this detailed analysis provides invaluable insights for policymakers and stakeholders, illuminating the complex dynamics of the composite index and supporting targeted interventions and improvements in critical areas for enhancing transport safety.

Decomposition of overall score change

The overall change in achievement scores for each country from 2010 to 2020 is illustrated in Fig. 10.

Fig. 10
figure 10

Transport safety achievement changes for each country from 2010 to 2020.

As depicted in Fig. 10, AG, BS, and UY exhibited the most significant advancements in transportation safety, followed by BB, AR, and BZ. In contrast, PE and VC remained unchanged. DM and SV experienced slight regressions, whereas KN, JM, LC, and GD displayed the most notable regressions.

To understand the drivers or sources of change in this holistic measure of achievement, we analyzed the factors or components that contributed to the changes in achievement between 2010 and 2020, as portrayed in Fig. 11. This analysis typically involves identifying and evaluating individual elements or sub-factors that influence the overall achievement and assessing their respective contributions to the observed changes.

Fig. 11
figure 11

Contributions of each SPIs to the overall change in safety score.

Figure 11 illustrates that indicators A1, A2, and A3 are critical for evaluating fatalities and transportation safety, providing policymakers essential insights to evaluate the impacts of previous transportation policies. Within the specified timeframe, A1 showed no contribution to any nation except VC, where it had a negative impact. Indicator A2 increased in six countries, with BS exhibiting the most significant improvement. Additionally, A3 contributed positively to only BR and was a non-contributing factor in other countries.

Indicators B1, B2, B3, and B4 are essential for assessing road user behavior, which significantly influences transportation safety. B1, which measures alcohol consumption—a major factor in traffic-related fatalities—improved only in AG, while KN and JM experienced the most significant declines. The adoption of seatbelts in both front and rear seats, and the use of motorcycle helmets, measured by B2, B3, and B4 respectively, are crucial for individual protection. A positive trend in adherence to B2, B3, and B4 was observed across twenty-five countries, although KN, GD, and JM displayed noticeable declines in these metrics, with PE exhibiting considerable regression in B3.

Indicators C1 and C2, which are related to safer road infrastructure, are vital for vehicle operation and overall transportation safety. Twenty-eight nations reported improvements in C1, which measures the percentage of paved roads, with notable progress in BS, AG, UY, and BB. Conversely, CR and LC experienced significant declines in this indicator. Regarding road network density, represented by C2, 11 countries enhanced their road network density, with AG, BB, and DM showing the most remarkable advancements, whereas KN, GD, and JM faced significant regressions.

Indicators D1 and D2 reflect the socioeconomic status of a nation and influence the government’s ability to allocate resources for transport safety infrastructure development. The urban population, indicated as D1, and GDP per capita, denoted as D2, increased in 29 states, with BS and AG experiencing the most significant growth. In contrast, KN exhibited the most notable decline in these metrics.

Indicators E1, E2, E3, and E4 focus on evaluating traffic policies and the enforcement of various transport safety laws. Twenty-seven states reported improvements across all four enforcement scores, including speed limit laws, drunk driving laws, seat belt laws, and helmet use laws, with the most significant enhancements observed in AG, UY, and BS. However, LC and GD experienced the most significant declines in E2, and DM in E3.

By analyzing changes in overall achievement, organizations and decision-makers can gain insights into the underlying factors driving these changes, prioritize areas for improvement or intervention, and make informed decisions to optimize outcomes.

Cross-country benchmarking within groups

To enhance knowledge sharing, collaboration, and continuous improvement within a group, it is crucial to conduct a comprehensive benchmarking of transport performance using specific indicators. This process promotes peer learning, adoption of best practices, and initiatives aimed at enhancing performance, thereby improving the overall performance of the group. Figures 12 and 13 illustrate the grouping of countries based on geographical distribution and the benchmarking of SPIs within each group, respectively.

Fig. 12
figure 12

Geographical location of the OAS countries concerning grouping.

Fig. 13
figure 13

International benchmarking within each group.

Figure 13 reveals that a top-ranking country within a group does not necessarily outperform other countries in all indicators. This variation highlights that each country possesses unique strengths and areas needing improvement, offering opportunities for mutual learning and information sharing.

In Group 1, Grenada (GD) shows a particularly strong score on indicator A3, while Barbados (BB) excels in C1 and C2. Trinidad and Tobago (TT) demonstrates notable strengths in E2 and E3, suggesting relative superiority in those aspects compared to its peers. Meanwhile, Saint Kitts and Nevis (KN) displays relatively consistent but moderate values across most indicators, without extreme highs or lows.

In Group 2, Canada (CA) and the United States (US) exhibit high scores in several categories, notably in behavioral and policy-related indicators such as B1 through B4, suggesting strong institutional frameworks or enforcement practices. Saint Vincent and the Grenadines (VC) stands out with top values in E3, indicating notable strengths in that dimension. In contrast, Chile (CL) and Argentina (AR) show more modest or varied performance across the indicators.

In Group 3, the Dominican Republic (DO) stands out with the highest score in A1 and a strong showing in E3, while Uruguay (UY) performs well in indicators E2, D1, and E1, suggesting relative strength in enforcement and institutional factors. Belize (BZ) shows elevated performance in B2 and B3, reflecting strengths in behavioral aspects. In contrast, Bolivia (BO) consistently shows lower values across most indicators, indicating areas for improvement. Mexico (MX) demonstrates relatively balanced outcomes with moderate values across several criteria, and Dominica (DM) shows peaks in A3 and B1.

In Group 4, Haiti (HT) demonstrates high values in A2 and E3, while Guyana (GY) performs strongly in A3 and B2, indicating strengths in enforcement and regulatory domains. Honduras (HN) and Nicaragua (NI) show elevated scores in E1 and E2, suggesting relatively strong institutional or operational aspects. In contrast, Guatemala (GT) and Peru (PE) exhibit more modest or dispersed performances across the indicators. Jamaica (JM) shows some strength in C1 and D1 but underperforms elsewhere.

In Group 5, Cuba (CU) shows high values in behavioral indicators such as B1 and B4, as well as strong scores in E1, suggesting effective safety practices and enforcement. Brazil (BR) stands out in E2 and D1, indicating robust institutional or infrastructural performance. Costa Rica (CR) and Colombia (CO) perform relatively well across multiple indicators, with Colombia showing higher values in A3 and B1. Paraguay (PY), while generally moderate, demonstrates consistent performance without extreme highs or lows.

In Group 6, Suriname (SR) exhibits high values in E2 and B1–B2, suggesting strengths in behavioral compliance and institutional support. El Salvador (SV) performs relatively well in E1 and E2, indicating robust enforcement mechanisms. Panama (PA) stands out in A3 and B1, reflecting strengths in strategic planning and behavior-related indicators. The Bahamas (BS) shows modest performance overall, with peaks in E1 and B4, while Venezuela (VE) scores highly in D1 and E3 but displays more variability across other indicators.

By conducting cross-country benchmarking within groups, stakeholders can obtain valuable insights into relative performance, identify strengths and weaknesses, promote knowledge exchange, and drive continuous improvement efforts among countries facing similar challenges and characteristics, supporting evidence-based policy refinement.

link

Leave a Reply

Your email address will not be published. Required fields are marked *