Conflict Dataset Research Articles

An object detection algorithm based on deep learning and salient feature fusion for roadside surveillance camera

AbstractIn intelligent transportation systems, object detection for a surveillance video is one of the important functions. The performance of existing surveillance video object detection algorithms is affected by the conflict between the features of the objects, which leads to a decline in precision. Therefore, an object detection algorithm based on deep learning and salient feature fusion is proposed. The proposed method introduces a non‐weight‐sharing network to process the salient features of the image and fuse them with the features extracted from the red blue green branch. Different from the previous solutions, the salient feature extraction branch uses the boundary features and statistical features of the image and fuses the features of the two branches in the efficient layer aggregation networks structure. At the same time, the attention module is used in efficient layer aggregation networks with convolutional block attention module to improve the efficiency of feature utilisation. The training and evaluation are carried out in the constructed surveillance video feature conflict dataset, and eight scenes are constructed in the way of orthogonal experiments. The experimental results show that the performance of object detection can be significantly improved by using the proposed method in the object detection task of the intelligent transportation system surveillance video feature conflict scene.

Sexual violence in the shadows: Private military contractors and the perpetration of sexual violence

How do private military contractors (PMCs) influence the perpetration of sexual violence? In providing the first empirical assessment of PMCs’ impact on conflict-related sexual violence (CRSV), I argue that, contrary to popular belief, PMC presence is negatively associated with the level of CRSV during armed conflict. Both states and PMCs have incentives to respect human rights. States prefer to avoid international condemnation while PMCs aim to be rehired. Association with human rights abuses damages both parties, leading each party to monitor the other to ensure good behavior. Furthermore, the global conflicts of the last 20 years have changed the face of warfare, producing professionalized warfighters and contractors. This raises the standards for PMCs, and results in more professionalism, especially in American contractors. Using the Private Military Contractors and Sexual Violence in Armed Conflict datasets from 1990 to 2008, I find PMC employment by state militaries is negatively associated with state-perpetrated CRSV. Additionally, both American contractors and contractors in general, employed in the Global War on Terror period are associated with lower levels of CRSV.

Demographic consequences of social movements: local protests delay marriage formation in Ethiopia

Abstract Despite their significance, life-course dynamics are rarely considered as consequences of social movements. We address this shortcoming by investigating the relationship between protest and marriage formation in Ethiopia. Building on scholarship in social movements and insights from family demography, we argue that exposure to protest delays marriage formation. To test our theoretical arguments, we created an original panel dataset using georeferenced data from the 2016 Ethiopia Demographic and Health Survey. We combined the marriage histories of 4,398 young women with fine-grained measures of exposure to local protests that we compiled from two conflict datasets covering events between 2002 and 2016. Using discrete-time event history analyses, we find that protest delays first-marriage formation. Additional analyses suggest that political uncertainty and disruptions in interethnic marriages cannot explain this effect. Instead, we provide tentative evidence that protest delays marriage formation by preoccupying large segments of the marriageable population, rendering them unavailable for this critical life-course transition. Our findings pave the way for scholarship on the demographic outcomes of protest and contribute to understanding marriage patterns in a country where the timing of marriage has far-reaching social consequences.

Sample Size Considerations for Fine-Tuning Large Language Models for Named Entity Recognition Tasks: Methodological Study.

Large language models (LLMs) have the potential to support promising new applications in health informatics. However, practical data on sample size considerations for fine-tuning LLMs to perform specific tasks in biomedical and health policy contexts are lacking. This study aims to evaluate sample size and sample selection techniques for fine-tuning LLMs to support improved named entity recognition (NER) for a custom data set of conflicts of interest disclosure statements. A random sample of 200 disclosure statements was prepared for annotation. All "PERSON" and "ORG" entities were identified by each of the 2 raters, and once appropriate agreement was established, the annotators independently annotated an additional 290 disclosure statements. From the 490 annotated documents, 2500 stratified random samples in different size ranges were drawn. The 2500 training set subsamples were used to fine-tune a selection of language models across 2 model architectures (Bidirectional Encoder Representations from Transformers [BERT] and Generative Pre-trained Transformer [GPT]) for improved NER, and multiple regression was used to assess the relationship between sample size (sentences), entity density (entities per sentence [EPS]), and trained model performance (F1-score). Additionally, single-predictor threshold regression models were used to evaluate the possibility of diminishing marginal returns from increased sample size or entity density. Fine-tuned models ranged in topline NER performance from F1-score=0.79 to F1-score=0.96 across architectures. Two-predictor multiple linear regression models were statistically significant with multiple R2 ranging from 0.6057 to 0.7896 (all P<.001). EPS and the number of sentences were significant predictors of F1-scores in all cases ( P<.001), except for the GPT-2_large model, where EPS was not a significant predictor (P=.184). Model thresholds indicate points of diminishing marginal return from increased training data set sample size measured by the number of sentences, with point estimates ranging from 439 sentences for RoBERTa_large to 527 sentences for GPT-2_large. Likewise, the threshold regression models indicate a diminishing marginal return for EPS with point estimates between 1.36 and 1.38. Relatively modest sample sizes can be used to fine-tune LLMs for NER tasks applied to biomedical text, and training data entity density should representatively approximate entity density in production data. Training data quality and a model architecture's intended use (text generation vs text processing or classification) may be as, or more, important as training data volume and model parameter size.

Are wars detrimental to the environment? Evidence from air pollution and land use

ABSTRACT This paper studies the long-run effect of war on environmental quality. Using data from the UCDP/PRIO Armed Conflict Dataset and the World Development Indicators, we apply a generalized difference-in-differences methodology. We compare the airborne emissions of countries that have experienced war after World War II (WWII) before and after the occurrence of the war event with the emissions of countries that did not experience any war after WWII. We find that, although wars decrease per capita CO2 emissions, they increase CO2 intensity. We also examine the effect of war on forest cover, which is found to be positive across all specifications. Lastly, we document differential effects by war type, country’s legal origins, and income level.

Introducing the Euro-Invasion Conflict Database 1513–1901

Abstract We introduce a comprehensive new dataset of conflicts, in chronological order, between the Indigenous Peoples of North America and Europeans seeking to colonize what is now the Continental United States. These data, covering 1,375 conflicts, were originally compiled as a book in 2007 by Michael L. Nunnally, a self-taught historian with a talent for careful data collection and analysis. A research team within the Rights Lab (University of Nottingham, U.K.), drew upon Nunnally’s American Indian Wars to code each recorded conflict (battle, skirmish, raid, massacre, etc.) along with other relevant information. The conflicts were then further researched by the Rights Lab team to add any missing information, resolve ambiguities or questions, and to find common conflict patterns through these centuries. The database holds thirty possible variables for each conflict. An example of one such variable is “genocidal massacre,” following the work of historian Benjamin Madley. Other variables include specifying and/or clarifying the identity of all participants; introducing, as often as possible, the exact or approximate GPS coordinates for each conflict; and noting who were the initiators of each conflict. The full Codebook is an Appendix to this article. To our knowledge, the database we are presenting is unique. We offer this dataset to scholars in the hope that it will allow for a more nuanced and deeper consideration of the history of the European invasion of North America over a four-hundred-year period.

Revolutionary days: Introducing the Latin American Guerrillas Dataset

The last two decades have witnessed an impressive expansion in the analysis of the causes, dynamics and legacies of internal armed conflicts. However, two relevant limitations persist. First, standard conflict datasets suffer from selection bias. And second, most available data lack a temporal approach that facilitates the analysis of insurgent lifecycles from the inception until the demobilization of non-state armed groups. To partially fill this gap, this article presents a novel dataset on revolutionary socialist insurgencies that were created in Latin America (LA) between 1950 and 2016. The Latin American Guerrillas Dataset (LAGD) covers the actions of 90 guerrilla organizations in 19 countries in the region, including annual level data on a variety of variables, such as level of success, political wings, rebel diplomacy, rebel governance, or number of combatants, among others. The LAGD represents one of the first systematic efforts for collecting comparative evidence on insurgencies operating in LA and should facilitate rigorous analyses on the divergent pathways of armed groups from a processual perspective, while dealing with extant issues of selection bias.

War Dummies: Structured Data on Organised Armed Confrontations with Dutch Involvement, 1566–1812

Abstract The War dummies dataset offers structured data on Dutch-involved organised armed confrontations from 1566 to 1812. Comprising 1216 records detailing the participation of 95 entities in 548 encounters across 50 wars, it fills a crucial need for well-structured, accessible, and reusable pre-1815 historical warfare data. Based on the comprehensive Military History of the Netherlands book series, it aligns with post-1815 conflict datasets like the Inter-State War Database of the Correlates of War Project and the Georeferenced Event Dataset of the Uppsala Conflict Data Program. This article outlines the data collection, structure, and potential research applications, and discusses data quality and potential biases. The War dummies codebook offers comprehensive variable descriptions.

Until consensus: Introducing the International Cyber Expression dataset

Extant scholarship has until now relied on informal-theoretic, case study, and interpretative methods to assess patterns of norm development in cyberspace. Ideally, these accounts would be complemented with more systematic cross-national and longitudinal empirical evidence. To address this gap, this article introduces the International Cyber Expression Dataset. The dataset includes a corpus of more than 34,000 official expressions of view by states and their authorized representatives regarding the international politics of cyberspace. The article describes the sources of these data and demonstrates the dataset’s usefulness, with an Online appendix containing an exploratory analysis of norm convergence. Future research can leverage the dataset to empirically test questions of theory and policy. For example, the dataset can be used to study how foundational theories of norm diffusion apply to cyberspace. It can also be paired with existing cyber conflict datasets to study the conditions under which state practice influences cyber discourse, and vice versa.

Livestock diversification mitigates the impact of farmer-herder conflicts on animal-source foods consumption in Nigeria

The escalation of farmer-herder conflicts poses a threat to agricultural production and livelihood outcomes in Nigeria. However, households with adaptive capacity may mitigate the negative impact of these conflicts on nutritious food consumption. In this study, we examine the impact of farmer-herder conflicts on animal-source foods (ASF) consumption and investigate the extent to which livestock diversification can serve as a mitigating factor. Using panel household data from Nigeria with a global georeferenced conflict dataset, we employ fixed-effects regression models to understand a causal relationship. Our findings reveal that exposure to farmer-herder conflicts reduces the quantity of ASF consumed and increases the number of days households exclude ASF from their diets. Additionally, we establish the role livestock diversification plays in mitigating the impacts of farmer-herder conflicts on ASF consumption. This evidence provides policymakers and practitioners with potential strategies for building nutrition resilience in locations that are exposed to farmer-herder conflicts. Promoting conflict-sensitive livestock production systems, such as cattle ranching, can be a strategy for sustaining nutrition and peacebuilding in Nigeria and countries in similar conflict situations.

Modeling and application of ship traffic conflicts in nearshore waters with heavy traffic

Ensuring the safety of ship navigation is essential to support the sustainable development of the shipping industry, and maritime safety analysis has thus received much attention in the field of shipping. Traditionally, the safety evaluation is mostly performed with the help of accident data, which is thus limited by the insufficient amount of accident data and affects the evaluation effect. In order to overcome this limitation, a ship traffic conflict model is established, which consists of three sub-models: multi-risk ship domain model, ship traffic conflict preliminary screening model and ship traffic conflict detection model. The multi-risk ship domain model is the key foundation, combining the traditional ship domain model with the water traffic conflict severity model to fully consider the severity of ship traffic conflict. To validate the proposed model, three ships in the study waters are selected as the own ships, and the severity of traffic conflicts formed with their target ships within a certain conflict time range are analyzed separately. Then, the conflict datasets obtained from the training of the ship traffic conflict model are used to analyze the traffic conflict situation in the heavily trafficked waters near Tianjin Port. Specifically, the spatial distribution characteristics are obtained using the kernel density estimation (KDE) method and the water grid method, and the temporal distribution characteristics are verified with the help of the accident statistics method. The proposed ship conflict model can provide technical support for further research in the field of ship traffic conflict, and the experimental analysis results are of great significance for strengthening maritime traffic supervision.

Bilateral Trade and Strategic Rivalry

Conflict is a costly endeavor. However, conflict itself is of unobservable magnitude which makes statistical inference problematic. The long-run economic cost of conflict is calculated as the sum of the contemporaneous costs and the discounted value of future costs. Typically, researchers use War or Militarized Interstate Conflicts as independent, discrete events to calculate its contemporaneous effect and then introduce a time binary variable to estimate the lagged effects since the end of the event. The conflict datasets accurately recognize the dates of the core conflict. However, they ignore the possibility that a lack of militarized conflict does not necessarily mean that issues have been settled, thus we are underestimating overall costs. The present study estimates the economic costs of rivalry. The international rivalry cycle is a process in which a pair of states create and sustain a relationship of atypical hostility for some period. This paper is part of the renaissance of research activity in the applied economics of international trade. The gravity model is used to determine the economic cost of Rivalry on bilateral trade using panel data. At the aggregate level, strategic and enduring rivalries have a negative and significant effect on trade flow. The results show that the total effect of rivalry accounts for 48%-57% of the fall in bilateral trade volume, which is equivalent in cost to 19% of the ad-valorem tax. If the rivalry is disaggregated by claim type: spatial, positional, and mixed, then we observe that the cost varies substantially with the type. Spatial rivalry explains 16%-26% of the fall in trade volume, while positional and mixed rivalry explain 49%-57% and 77%-82%, respectively.

Rebel group formation in Africa: Evidence from a new dataset

Theories of internal armed conflict onset often conceptualize an eruption emerging from pre-existing, non-violent opposition movements or organizations. This paper argues that this is not the most common pathway to armed conflict in Africa. It introduces new data on rebel group formation in 47 African states (1997–2015) that offers a more complete picture than existing conflict datasets; about two-thirds of the groups in the new dataset are not named in standard conflict datasets. The strong majority of these rebel groups formed in rural areas; did not emerge from nearby protest movements or civil society organizations; and initially did not benefit from substantial material resources nor commit large-scale violence. Analyses suggest that rebel groups forming in rural parts of African states draw less often on pre-war, public, non-violent contestation than those of urban areas. The paper highlights the need for more theory on the varied pathways to internal armed conflict, and the importance of – and challenges of collecting – systematic evidence on armed group formation.

Political instability patterns are obscured by conflict dataset scope conditions, sources, and coding choices

Conflict event datasets are used widely in academic, policymaking, and public spheres. Accounting for political violence across the world requires detailing conflict types, agents, characteristics, and source information. The public and policymaking communities may underestimate the impact of data collection decisions across global, real-time conflict event datasets. Here, we consider four widely used public datasets with global coverage and demonstrate how they differ by definitions of conflict, and which aspects of the information-sourcing processes they prioritize. First, we identify considerable disparities between automated conflict coding projects and researcher-led projects, largely resulting from few inclusion barriers and no data oversight. Second, we compare researcher-led datasets in greater detail. At the crux of their differences is whether a dataset prioritizes and mandates internal reliability by imposing initial conflict definitions on present events, or whether a dataset’s agenda is to capture an externally valid and comprehensive assessment of present violence patterns. Prioritizing reliability privileges specific forms of violence, despite the possibility that other forms actually occur; and leads to reliance on international and English-language information sources. Privileging validity requires a wide definition of political violence forms, and requires diverse, multi-lingual, and local sources. These conceptual, coding, and sourcing variations have significant implications for the use of these data in academic analysis and for practitioner responses to crisis and instability. These foundational differences mean that answers to “which country is most violent?”; “where are civilians most at risk?”; and “is the frequency of conflict increasing or decreasing?” vary according to datasets all purporting to capture the same phenomena of political violence.

Multicollinearity applied stepwise stochastic imputation: a large dataset imputation through correlation-based regression

This paper presents a stochastic imputation approach for large datasets using a correlation selection methodology when preferred commercial packages struggle to iterate due to numerical problems. A variable range-based guard rail modification is proposed that benefits the convergence rate of data elements while simultaneously providing increased confidence in the plausibility of the imputations. A large country conflict dataset motivates the search to impute missing values well over a common threshold of 20% missingness. The Multicollinearity Applied Stepwise Stochastic imputation methodology (MASS-impute) capitalizes on correlation between variables within the dataset and uses model residuals to estimate unknown values. Examination of the methodology provides insight toward choosing linear or nonlinear modeling terms. Tailorable tolerances exploit residual information to fit each data element. The methodology evaluation includes observing computation time, model fit, and the comparison of known values to replaced values created through imputation. Overall, the methodology provides useable and defendable results in imputing missing elements of a country conflict dataset.

Can the International Criminal Court prevent sexual violence in armed conflict?

Human rights activists, international organizations, and certain governments have championed prosecutions as a strategy to prevent conflict-related sexual violence, and the International Criminal Court (ICC) has focused extensively on prosecuting sexual violence crimes during its first two decades of operations. However, even as a growing body of empirical evidence suggests that the ICC contributes to improved human rights practices and the prevention of other atrocity crimes, such as violence against civilians, claims concerning the preventive effects of prosecutions on sexual violence remain largely untested. The purpose of this article is to test these claims. To this end, we analyzed the effects of ICC jurisdiction, interventions, and cross-case actions on sexual violence by government forces in intrastate conflicts from 1989 to 2018, using the Sexual Violence in Armed Conflict (SVAC) dataset. Contrary to the optimistic claims of proponents of prosecutions, we found that ICC jurisdiction and cross-case actions have negligible effects for this category of actors. We also found that ICC interventions are associated with increased sexual violence by government forces in intrastate conflicts. These findings suggest that prevention might require alternative—and, in some cases, potentially costlier—interventions.

Conflict Events Worldwide Since 1468BC: Introducing the Historical Conflict Event Dataset

Quantitative datasets of international conflict skew temporally to modern times and geographically and culturally to the West. Yet post–1815 conflicts featuring Western actors are only a small part of the history of warfare. Many scholars have bemoaned the potential selection bias which this introduces to studies of the causes and effects of military conflict, but as yet quantitative datasets which remedy both these temporal and geographic shortcomings have been lacking. Some datasets have expanded the scope of existing offerings temporally and others spatially, while others have attempted to expand both but with an important lack of detail in terms of location, participants, timing and outcomes. This dataset sets out to remedy the deficit. Using military history’s most extensive encyclopedia of conflict events, we have created a dataset of conflict events spanning the globe and a timescale from 1468BC to the invasion of Iraq in 2003, complete with precise geographic coordinates, year, participants and outcome. We demonstrate the promise of this data-set by using it to assess the frequently asserted relationship between conflict history and economic development, combined with Nordhaus’ GECON sub–national wealth data and historical data on population density from the Netherlands Environmental Agency.

The joint effects of terrorism and land access on livestock production decisions: Evidence from northern Nigeria

Livestock production is an integral part of the livelihoods of many households around the developing world and plays a significant role in farming households' food and nutrition security. However, conflict is a major challenge for livestock production in Africa and Nigeria in particular. We employ the Living Standards Measurement Study – Integrated Surveys on Agriculture (LSMS-ISA) panel data for Nigeria with a global georeferenced conflict dataset to examine the effect of terrorism on small-scale livestock production and the role of agricultural land access. Terrorism is an important factor that undermines livestock production. We find that increase in the incidents of terrorism reduces cattle herd size but does not reduce the herd size of households that manage a larger area of land. Also, terrorism significantly increases livestock diversification independently and jointly with land access. However, higher fatalities from terrorism reduces herd size irrespective of the size of land managed by households. Our findings suggest a plausible land abandonment in areas where terrorism is severe. Curbing terrorism in Nigeria would ensure farmers have physical access to their land and sustain livestock production.

Peace Negotiations in Civil Conflicts: A New Dataset

This article introduces the Peace Negotiations in Civil Conflicts (PNCC) dataset, which identifies whether a state-party and rival non-state armed group is in the formal negotiation phase of a peace process. The PNCC goes beyond recording instances of peace talks by offering a conceptual framework to identify when a government-rebel group dyad is at risk of formal negotiations. It explicitly considers issues central in event-history modelling, including censoring and the observation period. The PNCC also provides detailed information on negotiations, including the date and location of peace talks, and whether negotiations were bilateral or through mediation. The PNCC is the first source to distinguish mediated and non-mediated civil conflict negotiations under a single framework. Structured over the UCDP/PRIO Armed Conflict Dataset with global coverage for 1975-2013, the PNCC is integrable to commonly used civil war datasets.

The effectiveness in international mediation of international organizations

PurposeThis study aims to explore the factors whereby some international organizations (IOs) are more effective than others in international mediation and proposes three types of hypotheses through combining quantitative and qualitative analysis. First, IOs with greater institutional capabilities for gathering, exchanging and disseminating conflict-related information are more likely to mediate effectively. IO bias is another factor of influence in this regard. Second, IOs with greater institutional capabilities for deploying field missions and guaranteeing agreement are more likely to mediate effectively and maintain durable peace. Third, IOs with higher amounts of leverage are more likely to mediate effectively.Design/methodology/approachThe study establishes two data sets: one on interstate conflict; the other on intrastate conflict, thus to cover as many research samples as possible and avoid sampling bias.FindingsResults of the statistical analysis indicate that no matter interstate or intrastate conflict, IOs with higher institutional capabilities for diplomatic interventions are more likely to bring conflict parties to an agreement and thereafter maintain short-term peace. IOs with higher institutional capabilities for economic sanctions are similarly effective. Furthermore, IOs with greater institutional capabilities for field mission deployment mediate more effectively, whether in facilitating peace agreements or maintaining short-term and long-term peace after the agreement. IO bias and preference, however, have no significant impact on mediation effectiveness.Research limitations/implicationsThis study has made no in-depth explorations of such existing and important research areas as different third-party comparisons of the mediation effect.Practical implicationsThis paper attempts to make some contributions to the topic of mediation effectiveness through applying a bargaining model to the research and performing a statistical analysis based on both an interstate conflict data set and an intrastate conflict data set.Originality/valueThis paper provides an in-depth causal analysis and thoroughgoing comparison of the effectiveness of IOs in both interstate conflicts and intrastate conflicts.