Different Types Of Events Research Articles

Identifying geopolitical event precursors using attention-based LSTMs.

Forecasting societal events such as civil unrest, mass protests, and violent conflicts is a challenging problem with several important real-world applications in planning and policy making. While traditional forecasting approaches have typically relied on historical time series for generating such forecasts, recent research has focused on using open source surrogate data for more accurate and timely forecasts. Furthermore, leveraging such data can also help to identify precursors of those events that can be used to gain insights into the generated forecasts. The key challenge is to develop a unified framework for forecasting and precursor identification that can deal with missing historical data. Other challenges include sufficient flexibility in handling different types of events and providing interpretable representations of identified precursors. Although existing methods exhibit promising performance for predictive modeling in event detection, these models do not adequately address the above challenges. Here, we propose a unified framework based on an attention-based long short-term memory (LSTM) model to simultaneously forecast events with sequential text datasets as well as identify precursors at different granularity such as documents and document excerpts. The key idea is to leverage word context in sequential and time-stamped documents such as news articles and blogs for learning a rich set of precursors. We validate the proposed framework by conducting extensive experiments with two real-world datasets-military action and violent conflicts in the Middle East and mass protests in Latin America. Our results show that overall, the proposed approach generates more accurate forecasts compared to the existing state-of-the-art methods, while at the same time producing a rich set of precursors for the forecasted events.

Integrated drought monitoring and assessment using multi-sensor and multi-temporal earth observation datasets: a case study of two agriculture-dominated states of India.

In the current scenario of climate change, there has been a substantial increase in the frequency and severity of drought events. Therefore, it is necessary to investigate spatio-temporal characteristics of different drought events to plan for water resource utilization. The present study aims to assess and quantify the impact of meteorological, hydrological, and agricultural drought events from 2001 to 2017 over two large states of India (i.e., Maharashtra and Madhya Pradesh) using multi-temporal earth observation data at a finer resolution of 1km. Drought indices including Standardized Precipitation Index (SPI), Standardized Water level Index (SWI), and Vegetation Health Index (VHI) were derived from precipitation, groundwater level, vegetation indices, and land surface temperature data respectively to map the spatial extent and severity of meteorological, hydrological, and agricultural drought. Assessment of individual drought indices was carried out to understand the effect of these drought events separately on the study area. Area vulnerable with multiple droughts in the region was identified by integrating multiple drought indices to derive a composite drought map. This included the locations that are hotspots in terms of the occurrence of drought events of different types. The spatial pattern captured in the composite drought map indicates that most of the study areas are prone to drought events varying from mild to extreme severity. Madhya Pradesh is more prone to meteorological and agricultural drought events compared to hydrological drought. Maharashtra state is prone to three types of drought with agricultural drought being the dominant one. This study provides an opportunity to investigate and understand the drought phenomenon in a comprehensive manner at comparatively finer spatial resolution.

Real-time Accelerometer-Based Event Detection Through KNN Classification Method

Real-time pattern and activity recognition techniques have experienced a recent surge in the use of advanced statistical techniques for analyzing data. More explicitly, Machine Learning classification methods have been used in a variety of applications and for different purposes, such as recognizing different types of events from accelerometer data (e.g. in smart watches, wearables, and vibration analysis). The objective of this work is to present a novel use of accelerometer-based pattern recognition techniques to detect the urban space infrastructure characteristics based on their effect on a vehicle’s body frame dynamics. For this study, we focus on detecting speed bumps, potholes and curves based on a real-time data streaming with the objective of rendering a street digitalization to automatically update an urban space Digital Twin. The present study has achieved real-time event detection for an urban trajectory with high accuracy which yields data that can be later fused with other sensors data. This approach is effectively contributing to the infrastructure layer of a multi-layered approach of digital modelling. Furthermore, the results contribute to expand the body of knowledge of real-time accelerometer event recognition techniques.

Effects of emission sources on the particle number size distribution of ambient air in the residential area

Particle size distribution is a major factor in the health and climate effects of ambient aerosols, and it shows a large variation depending on the prevailing atmospheric emission sources. In this work, the particle number size distributions of ambient air were investigated at a suburban detached housing area in northern Helsinki, Finland, during a half-year period from winter to summer of 2020. The measurements were conducted with a scanning mobility particle sizer (SMPS) with a particle size range of 16–698 nm (mobility diameter), and the events with a dominant particle source were identified systematically from the data based on the time of the day and different particle physical and chemical properties. During the measurement period, four different types of events with a dominant contribution from either wood-burning (WB), traffic (TRA), secondary biogenic (BIO), or long-range transported (LRT) aerosol were observed. The particle size was the largest for the LRT events followed by BIO, WB, and TRA events with the geometric mean diameters of 72, 62, 57, and 41 nm, respectively. BIO and LRT produced the largest particle mode sizes followed by WB, and TRA with the modes of 69, 69, 46, and 25 nm, respectively. Each event type had also a noticeably different shape of the average number size distribution (NSD). In addition to the evaluation of NSDs representing different particle sources, also the effects of COVID-19 lockdown on specific aerosol properties were studied as during the measurement period the COVID-19 restrictions took place greatly reducing the traffic volumes in the Helsinki area in the spring of 2020. These restrictions had a significant contribution to reducing the concentrations of NOx and black carbon originating from fossil fuel combustion concentration, but insignificant effects on other studied variables such as number concentration and size distribution or particle mass concentrations (PM1, PM2.5, or PM10).

The Effects of Iconicity and Conventionalization on Word Order Preferences.

Of the six possible orderings of the three main constituents of language (subject, verb, and object), two-SOV and SVO-are predominant cross-linguistically. Previous research using the silent gesture paradigm in which hearing participants produce or respond to gestures without speech has shown that different factors such as reversibility, salience, and animacy can affect the preferences for different orders. Here, we test whether participants' preferences for orders that are conditioned on the semantics of the event change depending on (i) the iconicity of individual gestural elements and (ii) the prior knowledge of a conventional lexicon. Our findings demonstrate the same preference for semantically conditioned word order found in previous studies, specifically that SOV and SVO are preferred differentially for different types of events. We do not find that iconicity of individual gestures affects participants' ordering preferences; however, we do find that learning a lexicon leads to a stronger preference for SVO-like orders overall. Finally, we compare our findings from English speakers, using an SVO-dominant language, with data from speakers of an SOV-dominant language, Turkish. We find that, while learning a lexicon leads to an increase in SVO preference for both sets of participants, this effect is mediated by language background and event type, suggesting that an interplay of factors together determines preferences for different ordering patterns. Taken together, our results support a view of word order as a gradient phenomenon responding to multiple biases.

Multistate modeling and structure selection for multitype recurrent events and terminal event data.

In cardiovascular disease studies, a large number of risk factors are measured but it often remains unknown whether all of them are relevant variables and whether the impact of these variables is changing with time or remains constant. In addition, more than one kind of cardiovascular disease events can be observed in the same patient and events of different types are possibly correlated. It is expected that different kinds of events are associated with different covariates and the forms of covariate effects also vary between event types. To tackle these problems, we proposed a multistate modeling framework for the joint analysis of multitype recurrent events and terminal event. Model structure selection is performed to identify covariates with time-varying coefficients, time-independent coefficients, and null effects. This helps in understanding the disease process as it can detect relevant covariates and identify the temporal dynamics of the covariate effects. It also provides a more parsimonious model to achieve better risk prediction. The performance of the proposed model and selection method is evaluated in numerical studies and illustrated on a real dataset from the Atherosclerosis Risk in Communitiesstudy.

P140Cap Regulates the Composition and Localization of the NMDAR Complex in Synaptic Lipid Rafts.

The NMDARs are key players in both physiological and pathologic synaptic plasticity because of their involvement in many aspects of neuronal transmission as well as learning and memory. The contribution in these events of different types of GluN2A-interacting proteins is still unclear. The p140Cap scaffold protein acts as a hub for postsynaptic complexes relevant to psychiatric and neurologic disorders and regulates synaptic functions, such as the stabilization of mature dendritic spine, memory consolidation, LTP, and LTD. Here we demonstrate that p140Cap directly binds the GluN2A subunit of NMDAR and modulates GluN2A-associated molecular network. Indeed, in p140Cap KO male mice, GluN2A is less associated with PSD95 both in ex vivo synaptosomes and in cultured hippocampal neurons, and p140Cap expression in KO neurons can rescue GluN2A and PSD95 colocalization. p140Cap is crucial in the recruitment of GluN2A-containing NMDARs and, consequently, in regulating NMDARs' intrinsic properties. p140Cap is associated to synaptic lipid-raft (LR) and to soluble postsynaptic membranes, and GluN2A and PSD95 are less recruited into synaptic LR of p140Cap KO male mice. Gated-stimulated emission depletion microscopy on hippocampal neurons confirmed that p140Cap is required for embedding GluN2A clusters in LR in an activity-dependent fashion. In the synaptic compartment, p140Cap influences the association between GluN2A and PSD95 and modulates GluN2A enrichment into LR. Overall, such increase in these membrane domains rich in signaling molecules results in improved signal transduction efficiency.SIGNIFICANCE STATEMENT Here we originally show that the adaptor protein p140Cap directly binds the GluN2A subunit of NMDAR and modulates the GluN2A-associated molecular network. Moreover, we show, for the first time, that p140Cap also associates to synaptic lipid rafts and controls the selective recruitment of GluN2A and PSD95 to this specific compartment. Finally, gated-stimulated emission depletion microscopy on hippocampal neurons confirmed that p140Cap is required for embedding GluN2A clusters in lipid rafts in an activity-dependent fashion. Overall, our findings provide the molecular and functional dissection of p140Cap as a new active member of a highly dynamic synaptic network involved in memory consolidation, LTP, and LTD, which are known to be altered in neurologic and psychiatric disorders.

Maximum time between cardiac cycles in atrial fibrillation for assessing the risk of arterial thromboembolism

Aim. To assess the development of thromboembolic events in different types of atrial fibrillation (AF), which differ in the maximum time between cardiac cycles.Material and methods. The main group included 80 patients with permanent AF, while the control one — 88 people without AF. The following investigations were performed: echocardiography, 24-hour electrocardiographic (ECG) monitoring; ultrasound of extracranial arteries, aortic branches, lower limb arteries; sphygmography of the common carotid and posterior tibial arteries. If required, coronary angiography, cerebral angiography, brain computed tomography were performed. According to 24-hour ECG monitoring, all patients of the main group were divided into 2 subgroups (A and B) depending on the maximum time between cardiac cycles in AF: subgroup A (n=42) — patients with a maximum time between cardiac cycles <1,5 seconds, subgroup B (n=38) — ≥1,5. The followup period lasted 1 year. During the follow-up period, the development of arterial thromboembolic events was analyzed.Results. The patients were comparable in key paraclinical characteristics and comorbidity profile. One-year follow-up period revealed a significantly higher incidence of stroke, transient ischemic attack, myocardial infarction, and distal arterial embolism of lower limb arteries in subgroup B. With an increase in the maximum time between cardiac cycles in AF, an increase in hemodynamic parameters of arterial vessels occurred as both proximal and distal arteries. A similar trend was also observed in the analysis of arterial kinetic parameters.Conclusion. Not only the fact of AF presence is important for assessing the risk of arterial thromboembolism, but also its features. The most unfavorable for prognosis is AF with a maximum time between cardiac cycles ≥1,5 seconds. An increase in intra-arterial hemodynamic parameters after a long pause between ventricular contractions in AF without intracardiac thrombosis can become a key factor in the development of complications with existing plaques, which can become a source of distal embolism.

A Review of Event Deposits in Lake Sediments

Event deposits in lake sediments provide invaluable chronicles of geodynamic and climatic natural hazards on multi-millennial timescales. Sediment archives are particularly useful for reconstructing high-impact, low-frequency events, which are rarely observed in instrumental or historical data. However, attributing a trigger mechanism to event deposits observed in lake sediments can be particularly challenging as different types of events can produce deposits with very similar lithological characteristics, such as turbidites. In this review paper, we summarize the state of the art on event deposits in paleolimnology. We start by describing the sedimentary facies typical of floods, glacial lake outburst floods, avalanches, hurricanes, earthquakes, tsunamis, volcanic eruptions, and spontaneous delta collapses. We then describe the most indicative methods that can be applied at the scale of lake basins (geophysical survey, multiple coring) and on sediment cores (sedimentology, inorganic and organic geochemistry, biotic approach). Finally, we provide recommendations on how to obtain accurate chronologies on sediment cores containing event deposits, and ultimately date the events. Accurately identifying and dating event deposits has the potential to improve hazard assessments, particularly in terms of the return periods, recurrence patterns, and maximum magnitudes, which is one of the main geological challenges for sustainable worldwide development.

Correlations ofr-process elements in very metal-poor stars as clues to their nucleosynthesis sites

Aims.Various nucleosynthesis studies have pointed out that ther-process elements in very metal-poor (VMP) halo stars might have different origins. By means of familiar concepts from statistics (correlations, cluster analysis, and rank tests of elemental abundances), we look for causally correlated elemental abundance patterns and attempt to link them to astrophysical events. Some of these events produce ther-process elements jointly with iron, while others do not have any significant iron contribution. We try to (a) characterize these different types of events by their abundance patterns and (b) identify them among the existing set of suggestedr-process sites.Methods.The Pearson and Spearman correlation coefficients were used in order to investigate correlations amongr-process elements (X,Y) as well as their relation to iron (Fe) in VMP halo stars. We gradually tracked the evolution of those coefficients in terms of the element enrichments [X/Fe] or [X/Y] and the metallicity [Fe/H]. This approach, aided by cluster analysis to find different structures of abundance patterns and rank tests to identify whether several events contributed to the observed pattern, is new and provides deeper insights into the abundances of VMP stars.Results.In the early stage of our Galaxy, at least threer-process nucleosynthesis sites have been active. The first two produce and eject iron and the majority of the lighterr-process elements. We assign them to two different types of core-collapse events, not identical to regular core-collapse supernovae (CCSNe), which produce only light trans-Fe elements. The third category is characterized by a strongr-process and is responsible for the major fraction of the heavy mainr-process elements without a significant coproduction of Fe. It does not appear to be connected to CCSNe, in fact most of the Fe found in the relatedr-process enriched stars must come from previously occurring CCSNe. The existence of actinide boost stars indicates a further division among strongr-process sites. We assign these two strongr-process sites to neutron star mergers without fast black hole formation and to events where the ejecta are dominated by black hole accretion disk outflows. Indications from the lowest-metallicity stars hint at a connection with massive single stars (collapsars) forming black holes in the early Galaxy.

Risk factors for reactions and adverse effects during oral immunotherapy.

Oral immunotherapy (OIT) involves the potential for a variety of adverse events, which range from serious systemic reactions that require epinephrine to minimal oral reactions that require no treatment. This chapter describes common types of reactions seen in the course of OIT, reviews the frequency of and risk factors for different types of events as reported in recent literature (with a focus on real-world reports from private practice), and discusses treatment strategies for these adverse events. As the availability of OIT expands, it is paramount to ensure that allergists who offer OIT have a robust understanding of these reactions and mechanisms, with the overarching goal being the safety and tolerability of the therapy for the individual patient.

EVENT NETWORKS OF NON-FORMAL EDUCATION AS A RESULT OF SELF-ORGANIZATION OF ALL AGE COMMUNITIES

Introduction. The life of a modern person today is characterized by a constant renewal of the educational path, which allows one to quickly solve problems in both professional and personal development. The countries of Eurasia are no exception. The prioritized form of continuous education in the 21st century is non-formal education. It is democratic and makes exceptional use of the resources of civil society. Leading the development of non-formal education are self-organized communities of all ages, whose members create educational, cultural, and recreational activities in the network mode. In order to promote the educational activities of these communities, normative support is required, based on the provisions of the sciences of the interdisciplinary humanities field. The aim of the article. Describe the features of all age communities’ self-organization that form event networks. Research methods. Research methods include analyzing domestic and foreign studies, identifying scientific works whose results can be interpreted as possible solutions to the problems identified in this development, modeling methods, analysis and generalization of pedagogical experiences, pedagogical surveys, interviews, participant observation, and publicly speaking initiated and organized expertise. Results and Discussion. Rationale. To realize the developmental potential of non-formal practices in a formal educational setting, conditions must be created that attract a community of all ages (especially children and adults). Effective forms of self-organization of non-formal educational communities of all ages include the following: school communities for informal communication (CAS activities, music, sports, extracurricular academic subjects), families (family education and leisure), reading clubs, religious communities, travel groups, and city tour communities. This includes the active use of pedagogical practices of non-formal education: play and project activities, independent creative group activities, crowdsourcing, and participation in social networks. All age communities create networks of events of different types and forms, which make it possible to attract new members to the communities. Conclusion. In order to realize the development potential of non-formal education for all age communities, those responsible for formal educational institutions must create the conditions for the development of event networks: provide a venue and methodological support.

Link Prediction in Continuous-Time Dynamic Heterogeneous Graphs with Causality of Event Types

Dynamic heterogeneous graphs comprise different types of events with temporal labels. In many real-world scenarios, the temporal order of different types of events possibly implies causal relationships between these event types. However, existing methods designed to model dynamic heterogeneous graphs neglect the underlying causal relationships between event types. For instance, the determination of the occurrence of a new event is misled by irrelevant historical events considering the type and could lead to performance degradation. First, this paper explicitly defines the causality of event types by the heterogeneous causality graph to utilize such causality from the perspective of the graph structure to tackle the aforementioned issue. Second, this paper proposes the event type causality based continuous-time heterogeneous attention network (ECHN) to model dynamic heterogeneous graphs. ECHN aggregates features based on the strength of different causal relationships between event types in the prediction process to utilize the causality of event types from the perspective of the modeling algorithm. The utilities of event type causality weaken the negative effect of irrelevant events. Experimental results demonstrate that ECHN outperforms state-of-the-arts in the link prediction task. The authors believe that this paper is the first study to model the causality of event types in dynamic heterogeneous graphs explicitly.

A model of wild bee populations accounting for spatial heterogeneity and climate-induced temporal variability of food resources at the landscape level.

The viability of wild bee populations and the pollination services that they provide are driven by the availability of food resources during their activity period and within the surroundings of their nesting sites. Changes in climate and land use influence the availability of these resources and are major threats to declining bee populations. Because wild bees may be vulnerable to interactions between these threats, spatially explicit models of population dynamics that capture how bee populations jointly respond to land use at a landscape scale and weather are needed. Here, we developed a spatially and temporally explicit theoretical model of wild bee populations aiming for a middle ground between the existing mapping of visitation rates using foraging equations and more refined agent‐based modeling. The model is developed for Bombus sp. and captures within‐season colony dynamics. The model describes mechanistically foraging at the colony level and temporal population dynamics for an average colony at the landscape level. Stages in population dynamics are temperature‐dependent triggered by a theoretical generalized seasonal progression, which can be informed by growing degree days. The purpose of the LandscapePhenoBee model is to evaluate the impact of system changes and within‐season variability in resources on bee population sizes and crop visitation rates. In a simulation study, we used the model to evaluate the impact of the shortage of food resources in the landscape arising from extreme drought events in different types of landscapes (ranging from different proportions of semi‐natural habitats and early and late flowering crops) on bumblebee populations.

A strategy to identify event specific hospitalizations in large health claims databases

BackgroundHealth insurance claims data offer a unique opportunity to study disease distribution on a large scale. Challenges arise in the process of accurately analyzing these raw data. One important challenge to overcome is the accurate classification of study outcomes. For example, using claims data, there is no clear way of classifying hospitalizations due to a specific event. This is because of the inherent disjointedness and lack of context that typically come with raw claims data.MethodsIn this paper, we propose a framework for classifying hospitalizations due to a specific event. We then tested this framework in a private health insurance claims database (Symphony) with approximately 4 million US adults who tested positive with COVID-19 between March and December 2020. Our claims specific COVID-19 related hospitalizations proportion is then compared to nationally reported rates from the Centers for Disease Control by age.ResultsAcross all ages (18 +) the total percentage of Symphony patients who met our definition of hospitalized due to COVID-19 was 7.3% which was similar to the CDC’s estimate of 7.5%. By age group, defined by the CDC, our estimates vs. the CDC’s estimates were 18–49: 2.7% vs. 3%, 50–64: 8.2% vs. 9.2%, and 65 + : 14.6% vs. 28.1%.ConclusionsThe proposed methodology is a rigorous way to define event specific hospitalizations in claims data. This methodology can be extended to many different types of events and used on a variety of different types of claims databases.

An evaluation of public initiatives to change behaviours that affect water quality

AbstractPollution and environmental depletion are often caused by human behaviours, where if behaviours were modified, environmental pressure could be substantially reduced. Many public programmes aim to influence people to change their unsustainable behaviours but few undertake ex post evaluations of behavioural change programmes. This paper undertakes an evaluation of a 5‐year programme to understand whether community engagement activities lead to more sustainable practices. Using a treatment and control experiment, the research investigates whether programme participants are representative of the wider population, whether participation leads to sustainable practices, whether pro‐environmental behaviours are sustained over time, and the relative effectiveness of different types of events on individual behaviours. Overall, results suggest that water engagement event participants are more likely to adopt sustainable behaviours, however attendees only represent a small cohort of the wider population, which possibly hinders a wider adoption of good practices. With respect to individual behaviours, different outcomes were detected for different behaviours, which suggests that not all behaviours were equally amenable to change. When comparing the impact of different events, events that are focused on building community ties were more successful than events with simple provision of information. Finally, new pro‐environmental behaviours tend to be abandoned after a period of time, therefore re‐engagement at regular time intervals is advised.

Atrial fibrillation: the difference in thromboembolic events prognosis

Abstract Funding Acknowledgements Type of funding sources: None. Aim To identify the difference between the thromboembolic events in different types of atrial fibrillation (AF). Materials and methods We included 188 patients with permanent AF (of them were 99 men and 89 women). According to the data of 24-hours ECG monitoring, all patients were divided into 3 groups depending on the duration of maximum pauses between ventricular complexes in AF: 1. Patients with a pause < 1 second (64 people). 2. Patients with a pause of ≥1, but <2 seconds (62 people). 3. Patients with a pause ≥ 2 seconds or more (62 people). In control group we included 80 patients without AF. All groups were identical in age, sex and comorbidities. All patients underwent transthoracic or transesophageal echocardiography (we excluded the intra-chamber thrombosis in all the patients as well as the left ventricle aneurisms), Doppler ultrasound of the brachiocephalic vessels, branches of the aorta, arteries of the lower extremities, renal arteries, lipids level. If it was indicated, we performed stress echocardiography, coronary angiography, renal artery angiography, computed tomography of brain with contrast. The patients of all groups were on the standard therapy recommended by current ESC guidelines. We were observing the patients for 1 year after the beginning of the investigation. The endpoints were at 3, 6 and 12 months after the primary visit. During these endpoints, we analyzed the development of thromboembolic complications. Results We analyzed the data obtained on thromboembolic complications in 1 year in patients of 1, 2, 3 groups and compared them with the control group. The most frequent incidence of complications was in group 3 compared with other groups and the control group. We believe that the risk factor for thromboembolic complications is not only the fact of permanent form of AF, but also what type of AF. The most unfavorable is AF with pauses between ventricular contractions of 2 seconds or more. Conclusion We consider to use the proposed functional classification of AF in clinical practice. Determination of the type of AF will largely determine the further management tactics of each patient, the prognosis of the thromboembolic complications. We supplemented the CHA2DS2-VASc score with the new independent risk factor – type of AF – in accordance to the maximum duration of pauses between cardiocycles.

Mathematical modeling of free convective MHD flow of temperature and heap transfer effects on caisson fluid over an oscillatory vertical plate

In general, some static analysis of the predictable flow of the compressible and caisson fluid of some infinite vertical plates is determined by taking into account its varying mass or constant fluid flow and the uniform increase of the partition temperature. Differential equations can be predicted using different components of the Laplace transformation technique. Usually, two different solutions are available when the fluid flows in a particular direction. In this paper, an improved method for calculating the velocity of different fluids is proposed. Its possibilities are designed to calculate the different types of events that take place near the high temperature plates of the liquid. The results are compared with the flow along the plates at a near constant temperature.

Mathematical modeling of free convective MHD flow of temperature and heap transfer effects on caisson fluid over an oscillatory vertical plate

In general, some static analysis of the predictable flow of the compressible and caisson fluid of some infinite vertical plates is determined by taking into account its varying mass or constant fluid flow and the uniform increase of the partition temperature. Differential equations can be predicted using different components of the Laplace transformation technique. Usually, two different solutions are available when the fluid flows in a particular direction. In this paper, an improved method for calculating the velocity of different fluids is proposed. Its possibilities are designed to calculate the different types of events that take place near the high temperature plates of the liquid. The results are compared with the flow along the plates at a near constant temperature.

Matching and analysing conservation–restoration trajectories

The context of this work is an on-going project at the French National Library (BnF), which aims at providing predictions of the documents physical state based on their conservation–restoration histories. A document can be either in a good state and available to the readers, or damaged and unavailable to them. As libraries may contain millions of documents, the manual monitoring and analysis of their physical state is not realistic in practice. We therefore propose to analyse their conservation histories in order to derive reliable predictions of their physical state. To achieve this goal, we introduce in this paper the following contributions. First, we propose a representation of a document conservation history as a conservation–restoration trajectory, and we define its different types of events. We also propose a trajectory matching process that computes a similarity score between two conservation–restoration trajectories considering the terminological heterogeneity of the events, using an ontological model that represents the domain experts knowledge. Second, we provide a trajectory analysis process which identifies the most representative sequences of events of the deteriorated documents. Finally, we propose a prediction model for the physical state of the documents based on the trajectory analysis process. We present some experiments showing the effectiveness of the matching process as well as the prediction model.