Co-occurrence Model Research Articles

Co-Existence Ecology of Large Predators in Belize

Jaguars, pumas, and ocelots are the main large predators within Belize, Central America that co-exist together and compete for resources. This study focused on determining whether the occurrence and/or detection of smaller predators (ocelots and pumas) is influenced by larger, more dominant predators (jaguars) across the landscape. Additionally, within this study we examined the influence of habitat-type and individual species sex on co-existence ecology utilizing data collected from non-invasive trail cameras in Belize. We formatted raw photo data into “capture histories” for three target species at each camera station across four study sites, and calculated the species interaction factors (SIFs) through co-occurrence modeling in program PRESENCE. Data analysis is currently on-going, but we hypothesize that the top-down control hypothesis, which states top predators will negatively influence space use and/or detection by smaller predators. We expect this will lead to negative SIFs in pairwise comparison indicating avoidance of top predators by smaller ones. Within each species, our preliminary results for the first of four study show that males tend to occur at a camera station shortly after females. Thus we expect this to lead to positive SIF values (indicating attraction) between the sexes of the same species. While analyses are not yet complete, we expect our results to give insight into species co-existence among competitors and also provide a greater understanding to whether sex and/or habitat influences competing species interactions.

Co-occurrence of snow leopard Panthera uncia, Siberian ibex Capra sibirica and livestock: potential relationships and effects

AbstractUnderstanding the impact of livestock on native wildlife is of increasing conservation relevance. For the Vulnerable snow leopard Panthera uncia, wild prey reduction, intensifying human–wildlife conflicts and retaliatory killings are severe threats potentially exacerbated by the presence of livestock. Elucidating patterns of co-occurrence of snow leopards, wild ungulate prey, and livestock, can be used to assess the compatibility of pastoralism with conservation. We used camera trapping to study the interactions of livestock, Siberian ibex Capra sibirica and snow leopards in a national park in the Altai mountains, Mongolia. We obtained 494 detections of wild mammals and 912 of domestic ungulates, dogs and humans. Snow leopards and Siberian ibex were recorded 14 and 33 times, respectively. Co-occurrence modelling showed that livestock had a higher estimated occupancy (0.65) than ibex, whose occupancy was lower in the presence of livestock (0.11) than in its absence (0.34–0.35 depending on scenarios modelled). Snow leopard occupancy did not appear to be affected by the presence of livestock or ibex but the robustness of such inference was limited by uncertainty around the estimates. Although our sampling at presumed snow leopard passing sites may have led to fewer ibex detections, results indicate that livestock may displace wild ungulates, but may not directly affect the occurrence of snow leopards. Snow leopards could still be threatened by livestock, as overstocking can trigger human–carnivore conflicts and hamper the conservation of large carnivores. Further research is needed to assess the generality and strength of our results.

Temporal dynamics and longitudinal co-occurrence of depression and different anxiety syndromes in youth: Evidence for reciprocal patterns in a 3-year prospective study

BackgroundDepression is highly comorbid with anxiety in youth. It is frequently reported that anxiety precedes depression; however, evidence surrounding the temporal precedence of anxiety over depression is mixed. Many studies of anxiety-depression co-occurrence lump distinct forms of anxiety, obscuring information regarding trajectories of specific anxiety syndromes. This study sought to more accurately describe the development of anxiety and depression over time by moving beyond the question of temporal precedence to investigate a developmentally dynamic model of anxiety-depression co-occurrence. MethodsA community sample of 665 youth (M= 11.8, SD= 2.4; 55% female) completed repeated self-report measures of depression and anxiety (social, physical, and separation anxiety) over a 3-year longitudinal study. Prospective associations between distinct syndromes of anxiety with depression were analyzed using an autoregressive cross-lagged path model over four time points. ResultsPhysical symptoms and depression symptoms reciprocally predicted each other, above and beyond the stability of either domain. Social anxiety and depression symptoms similarly predicted each other in a systematic pattern. LimitationsOur study is limited in its generalizability to other forms of anxiety, like worry. Additional research is needed to determine whether similar patterns exist in clinical populations, and whether these processes maintain symptoms once they reach diagnostic levels. ConclusionsThe development of syndromes of depression, physical, and social anxiety during childhood and adolescence occurs in a predictable, systematic reciprocal pattern, rather than sequentially and unidirectionally (i.e., anxiety syndromes precede depression). Results are clinically useful for predicting risk for disorder, and demonstrate the necessity of tracking symptom levels across domains.

Ant mosaics in Bornean primary rain forest high canopy depend on spatial scale, time of day, and sampling method.

BackgroundCompetitive interactions in biological communities can be thought of as giving rise to “assembly rules” that dictate the species that are able to co-exist. Ant communities in tropical canopies often display a particular pattern, an “ant mosaic”, in which competition between dominant ant species results in a patchwork of mutually exclusive territories. Although ant mosaics have been well-documented in plantation landscapes, their presence in pristine tropical forests remained contentious until recently. Here we assess presence of ant mosaics in a hitherto under-investigated forest stratum, the emergent trees of the high canopy in primary tropical rain forest, and explore how the strength of any ant mosaics is affected by spatial scale, time of day, and sampling method.MethodsTo test whether these factors might impact the detection of ant mosaics in pristine habitats, we sampled ant communities from emergent trees, which rise above the highest canopy layers in lowland dipterocarp rain forests in North Borneo (38.8–60.2 m), using both baiting and insecticide fogging. Critically, we restricted sampling to only the canopy of each focal tree. For baiting, we carried out sampling during both the day and the night. We used null models of species co-occurrence to assess patterns of segregation at within-tree and between-tree scales.ResultsThe numerically dominant ant species on the emergent trees sampled formed a diverse community, with differences in the identity of dominant species between times of day and sampling methods. Between trees, we found patterns of ant species segregation consistent with the existence of ant mosaics using both methods. Within trees, fogged ants were segregated, while baited ants were segregated only at night.DiscussionWe conclude that ant mosaics are present within the emergent trees of the high canopy of tropical rain forest in Malaysian Borneo, and that sampling technique, spatial scale, and time of day interact to determine observed patterns of segregation. Restricting sampling to only emergent trees reveals segregatory patterns not observed in ground-based studies, confirming previous observations of stronger segregation with increasing height in the canopy.

Translocation Breakpoints Preferentially Occur in Euchromatin and Acrocentric Chromosomes.

Chromosomal translocations drive the development of many hematological and some solid cancers. Several factors have been identified to explain the non-random occurrence of translocation breakpoints in the genome. These include chromatin density, gene density and CCCTC-binding factor (CTCF)/cohesin binding site density. However, such factors are at least partially interdependent. Using 13,844 and 1563 karyotypes from human blood and solid cancers, respectively, our multiple regression analysis only identified chromatin density as the primary statistically significant predictor. Specifically, translocation breakpoints preferentially occur in open chromatin. Also, blood and solid tumors show markedly distinct translocation signatures. Strikingly, translocation breakpoints occur significantly more frequently in acrocentric chromosomes than in non-acrocentric chromosomes. Thus, translocations are probably often generated around nucleoli in the inner nucleoplasm, away from the nuclear envelope. Importantly, our findings remain true both in multivariate analyses and after removal of highly recurrent translocations. Finally, we applied pairwise probabilistic co-occurrence modeling. In addition to well-known highly prevalent translocations, such as those resulting in BCR-ABL1 (BCR-ABL) and RUNX1-RUNX1T1 (AML1-ETO) fusion genes, we identified significantly underrepresented translocations with putative fusion genes, which are probably subject to strong negative selection during tumor evolution. Taken together, our findings provide novel insights into the generation and selection of translocations during cancer development.

TWCM: Twitter Word Co-occurrence Model for Event Detection

Increasing popularity of Twitter data indicates that social media is up-to-date source of information in terms of recent topics, trends or events. Although many existing techniques provide useful insights from social media data, but word co-occurrence structure based analysis is still unexplored. A word co-occurrence network based event detection approach has been proposed in this research work. The proposed technique is independent of static parameters and is an unsupervised technique. The proposed approach use directed and weighted graph for analysis and topological sorting for identifying event phrases. The proposed approach is validated on the First Story Detection standard dataset. The results are compared with the existing techniques namely Frequent Pattern Mining using Dynamic Support Values (FPM-DSV), Soft Frequent Pattern Mining (SFPM), High utility pattern clustering (HUPC) and Transaction-based Rule Change Mining (TRCM). The results show that the proposed technique outperforms FPM-DSV, SFPM, HUPC and TRCM in terms of recall rate and redundancy rate.

Activity patterns and spatial co-occurrence of sympatric mammals in the moist temperate forest of the Kashmir Himalaya, India

Little is known about the activity pattern and co-occurrence of mammalian species coexisting sympatrically in the rugged terrains of the Himalaya. By using camera traps we quantified diel activity patterns and temporal overlap as well as probabilistic model of spatial species co-occurrence amongst eight species. The jungle cat and the Indian porcupine tended to be strictly nocturnal whereas the leopard cat and the red fox were primarily nocturnal and crepuscular, yellow-throated marten was diurnal. The Asiatic black bear, leopard and the hangul showed no clear pattern. We found highest temporal overlap (Δ1 = 0.84, 95%CI = 0.54-0.86) between leopard and Asiatic black bear followed by pairs between red fox and leopard cat (Δ1 = 0.83, 95%CI = 0.72-0.91) and jungle cat (Δ1 = 0.82, 95%CI = 0.60-0.91). The lowest temporal overlap was found between yellow-throated marten and Indian porcupine (Δ = 0.11, 95%CI = 0.08-0.20). The results from probabilistic pairwise species co-occurrence analysis suggest that out of 28 possible species pairs 25 had random associations and one pair was unclassifiable. The remaining two species pairs had positive associations of co-occurrences between leopard cat and hangul (Pgt < 0.05) and Indian porcupine and hangul (Pgt < 0.05). Our results provided new insights into this unique community of mammals of the western Himalaya and will facilitate future studies on the mechanism determining coexistence of animal species within complex system.

When habitat matters: Habitat preferences can modulate co-occurrence patterns of similar sympatric species.

Disentangling the role of competition in regulating the distribution of sympatric species can be difficult because species can have different habitat preferences or time use that introduce non-random patterns that are not related to interspecific interactions. We adopted a multi-step approach to systematically incorporate habitat preferences while investigating the co-occurrence of two presumed competitors, morphologically similar, and closely related ground-dwelling birds: the brown tinamou (Crypturellus obsoletus) and the tataupa tinamou (C. tataupa). First, we used single-species occupancy models to identify the main landscape characteristics affecting site occupancy, while accounting for detection probability. We then used these factors to control for the effect of habitat while investigating species co-occurrence. In addition, we investigated species present-time partitioning by measuring the degree of overlap in their activity time. Both species were strictly diurnal and their activity time highly overlapped (i.e., the species are not present-time partitioning). The distribution of the two species varied across the landscape, and they seemed to occupy opposite portions of the study area, but co-occurrence models and species interaction factors suggested that the tinamous have independent occupancy and detection. In addition, co-occurrence models that accounted for habitat performed better than models without habitat covariates. The observed co-occurrence pattern is more likely related to habitat preferences, wherein species segregated by elevation. These results provide evidence that habitat characteristics can play a bigger role than interspecific interactions in regulating co-existence of some species. Therefore, exploring habitat preferences while analyzing co-occurrence patterns is essential, in addition to being a feasible approach to achieve more accurate estimation of parameters reflecting species interactions. Occupancy models can be a valuable tool in such modeling.

Threats to a rainforest carnivore community: A multi-year assessment of occupancy and co-occurrence in Madagascar

Protected areas (PA) aim to eliminate many of the threats that species face on the greater landscape. In the last three decades, PA's have expanded considerably; however, quantitative assessments of how well they have mitigated threats to habitat and biodiversity are very limited. Habitat bordering PA's and the wildlife that use it are threatened by a wide-range of anthropogenic pressures (e.g., edge effects, fragmentation, and introduced predators) and this situation is particularly acute for low-density, poorly studied carnivore communities. From 2010 to 2015, we photographically sampled within (contiguous forest) and bordering (degraded, fragmented forest) a UNESCO World Heritage rainforest PA in Madagascar - Ranomafana National Park (RNP). We investigated the effects of invasive predators, local people presence, and habitat quality on the endemic rainforest carnivore community using static, dynamic, and co-occurrence models. We found native carnivores to be absent or have a low probability of occurrence in degraded forest bordering the PA, while local people and dogs (Canis familiaris) had high occurrence. Madagascar's largest endemic carnivore, the fosa (Cryptoprocta ferox) and the much smaller ring-tailed vontsira (Galidia elegans), occurrence in RNP declined rapidly over six years; their strong co-occurrence with dogs suggests interspecific competition, direct aggression/mortality, or disease as the cause. We highlight the dangers posed to biodiversity, particularly carnivores, from anthropogenic pressures bordering PA's and present recommendations to address increased human and dog activity, including programs to control dogs and their impact on biodiversity.

Formal Models of the Network Co-occurrence Underlying Mental Operations.

Systems neuroscience has identified a set of canonical large-scale networks in humans. These have predominantly been characterized by resting-state analyses of the task-unconstrained, mind-wandering brain. Their explicit relationship to defined task performance is largely unknown and remains challenging. The present work contributes a multivariate statistical learning approach that can extract the major brain networks and quantify their configuration during various psychological tasks. The method is validated in two extensive datasets (n = 500 and n = 81) by model-based generation of synthetic activity maps from recombination of shared network topographies. To study a use case, we formally revisited the poorly understood difference between neural activity underlying idling versus goal-directed behavior. We demonstrate that task-specific neural activity patterns can be explained by plausible combinations of resting-state networks. The possibility of decomposing a mental task into the relative contributions of major brain networks, the "network co-occurrence architecture" of a given task, opens an alternative access to the neural substrates of human cognition.

Crustacean communities in coastal ephemeral pools in the Araucanía region (38° S, Chile).

The fauna communities of ephemeral pools in southern Chile are characterized by heterogeneity of crustacean taxa; nevertheless, no detailed studies exist of their community structure. The aim of the present study was to analyze the crustacean community structure in two groups of ephemeral pools (Puaucho and Nigue pools) in the coastal zone of the Araucanía region. A correlation matrix was made by species abundance against temperature, conductivity, pH and total dissolved solids. In a second step, a null model for species co-occurrence was applied to the total data and to each group. The results for total data revealed a significant direct relation between the abundance of H. costera, C. dubia and Mesocyclops. For the Puaucho pools, the same results were found together with direct associations with total dissolved solids, conductivity and pH. Finally, different results were found for the Nigue pools, with no clear significant associations, either direct or indirect, between the abundance of different crustacean taxa and abiotic parameters. These results were supported by the co-occurrence null model analysis, which revealed the presence of regulator factors for the total data, and for each of the two groups. Ecological topics are discussed with emphasis on meta-community dynamics.

The principals of meaning: Extracting semantic dimensions from co-occurrence models of semantics.

Notable progress has been made recently on computational models of semantics using vector representations for word meaning (Mikolov, Chen, Corrado, & Dean, 2013; Mikolov, Sutskever, Chen, Corrado, & Dean, 2013). As representations of meaning, recent models presumably hone in on plausible organizational principles for meaning. We performed an analysis on the organization of the skip-gram model's semantic space. Consistent with human performance (Osgood, Suci, & Tannenbaum, 1957), the skip-gram model primarily relies on affective distinctions to organize meaning. We showed that the skip-gram model accounts for unique variance in behavioral measures of lexical access above and beyond that accounted for by affective and lexical measures. We also raised the possibility that word frequency predicts behavioral measures of lexical access due to the fact that word use is organized by semantics. Deconstruction of the semantic representations in semantic models has the potential to reveal organizing principles of human semantics.

Category co-occurrence modeling for large scale scene recognition

Scene recognition involves complex reasoning from low-level local features to high-level scene categories. The large semantic gap motivates that most methods model scenes resorting to mid-level representations (e.g. objects, topics). However, this implies an additional mid-level vocabulary and has implications in training and inference. In contrast, the semantic multinomial (SMN) represents patches directly in the scene-level semantic space, which leads to ambiguity when aggregated to a global image representation. Fortunately, this ambiguity appears in the form of scene category co-occurrences which can be modeled a posteriori with a classifier. In this paper we observe that these patterns are essentially local rather than global, sparse, and consistent across SMNs obtained from multiple visual features. We propose a co-occurrence modeling framework where we exploit all these patterns jointly in a common semantic space, combining both supervised and unsupervised learning. Based on this framework we can integrate multiple features and design embeddings for large scale recognition directly in the scene-level space. Finally, we use the co-occurrence modeling framework to develop new scene representations, which experiments show that outperform previous SMN-based representations.

Using codispersion analysis to characterize spatial patterns in species co-occurrences.

Visualizing and quantifying spatial patterns of co-occurrence (i.e., of two or more species, or of species and underlying environmental variables) can suggest hypotheses about processes that structure species assemblages and their relevant spatial scales. Statistical models of spatial co-occurrence generally assume that underlying spatial processes are isotropic and stationary, but many ecologically realistic spatial processes are anisotropic and non-stationary. Here, we introduce codispersion analysis to ecologists and use it to detect and quantify anisotropic and nonstationary patterns and their relevant spatial scales in bivariate co-occurrence data. Simulated data illustrated that codispersion analysis can accurately characterize complex spatial patterns. Analysis of co-occurrence of common tree species growing in a 35-ha plot revealed both positive and negative codispersion between different species; positive codispersion values reflected positive correlation in species abundance (aggregation), whereas negative codispersion values reflected negative correlation in species abundance (segregation). Comparisons of observed patterns with those simulated using two different null models showed that the codispersion of most species pairs differed significantly from random expectation. We conclude that codispersion analysis can be a useful exploratory tool to guide ecologists interested in modeling spatial processes.

Effects of Prey Presence and Scale on Bobcat Resource Selection during Winter.

Factors relevant to resource selection in carnivores may vary across spatial and temporal scales, both in magnitude and rank. Understanding relationships among carnivore occupancy, prey presence, and habitat characteristics, as well as their interactions across multiple scales, is necessary to improve our understanding of resource selection and predict population changes. We used a multi-scale dynamic hierarchical co-occurrence model with camera data to study bobcat and snowshoe hare occupancy in the Upper Peninsula of Michigan during winter 2012–2013. Bobcat presence was influenced at the local scale by snowshoe hare presence, and by road density at the local and larger scale when hare were absent. Hare distribution was related primarily to vegetation cover types, and detectability varied in space and time. Bobcat occupancy dynamics were influenced by different factors depending on the spatial scale considered and the resource availability context. Moreover, considering observed co-occurrence, we suggest that bobcat presence had a greater effect on hare occupancy than hare presence on bobcat occupancy. Our results highlight the importance of studying carnivore distributions in the context of predator-prey relationships and its interactions with environmental covariates at multiple spatial scales. Our approach can be applied to other carnivore species to provide insights beneficial for management and conservation.

Showing their true colours: Possible secular declines and a Jensen effect on colour acuity — More evidence for the weaker variant of Spearman's Other Hypothesis

ABSTRACT Spearman's Other Hypothesis predicts that the common factor amongst sensory discrimination measures corresponds to general intelligence (g). The co-occurrence model predicts that low-complexity physiological information-processing indicators reliably measure g across cohorts, and should therefore decline with time due to genetic changes in the broader population. As strong relations exist between general sensory discrimination and g, such measures should show evidence of secular declines. This is tested using N-weighted temporal regression of square-root Total Error Scores (√ TES), obtained from four Western normative samples published in the 1980s, 90s and 2000s (combined N = 752) evaluated using the Farnsworth–Munsell 100-Hue colour acuity test (disattenuated g loading = .78). A significant temporal β value of .37 was found (controlling for national IQ), suggesting a decline in colour acuity equating to a reduction in g of − 3.15 points per decade. Analysis of the subset of the cohorts aged 20–29 years, in which colour acuity is maximized, reveals a larger secular decline (β = .67, N = 199, − 5.85 points per decade). The small number of studies employed in these analyses makes these findings tentative however. Also consistent with a weaker variant of the Other Hypothesis is the finding that 100-Hue acuity-IQ correlations are associated with the Jensen effect. The aggregate vector correlation across two studies is .63 (N = 932.5, p

Patterns of spatial co-occurrence among native and exotic carnivores in north-eastern Madagascar

Human populations continue to increase and encroach on remaining natural habitats worldwide, resulting in greater numbers and larger ranges of commensal exotic carnivores such as cats and dogs. This results in increased interactions with native wildlife. In Madagascar, we know relatively little about the effects of domestic and/or feral dogs and cats on native carnivore populations. We investigated spatial interactions by combining photographic sampling across seven sites with two-species co-occurrence modeling to provide the first assessment of the spatial co-occurrence of native and exotic carnivores in Madagascar, including an examination of habitat characteristics that explain these relationships. Our surveys from 2008 to 2013 accumulated 2991 photo-captures of native and exotic carnivores in 8854 trap nights. We found that native and exotic carnivores in rainforest habitat occur together less often than expected and that exotic carnivores may be replacing native carnivores, particularly in forest areas nearest villages. Six of the native carnivores in this study had higher site use in the absence of exotic carnivores and their species interaction factors (SIF) revealed a lack of co-occurrence (e.g. SIF < 1.0). We found that nocturnal and/or crepuscular native carnivores were less likely to co-occur with exotic carnivores. We demonstrate the effectiveness of combining photographic sampling with co-occurrence modeling to investigate the effects of exotic carnivores on an entire community of native carnivores. Our study exposes the strong negative influence of exotic carnivores, ranging from exclusion to complete replacement of native carnivores, and we urgently recommend a combination of targeted educational programs and removal programs to combat the influx of exotic carnivores.

Ecological diversity, community structure and conservation of Niger Delta mammals

The Niger Delta (southern Nigeria) is a crucial region for biodiversity because it is (i) part of the West African forests hotspot, (ii) under considerable threats because of habitat devastations, and (iii) a minor centre of endemism for many faunal and floral groups. In this paper, we analyze the communities of mammals (excluding bats, shrews and murids) occurring in all vegetation zones of the region, using both original field data (direct sightings of live animals, bushmeat market specimens, skins, skulls, bones, tracks, reliable interviews) collected in 1996–2015 and literature data. We also explore the main drivers of community structure by using a suite of multivariate (cluster) and null models (co-occurrence) analyses, and the conservation consequences of our data. A total of 45 extant species was found; their ecological distribution was uneven across vegetation zones, with flood forest, marsh forest and eastern flank forests housing more species. Although most species were widespread, 33.3 % had narrow ecological preferences. Cluster analysis confirmed that the community composition was (i) similar in habitats that were suspected to derive from pristine flood forests, and that (ii) the mammal assemblages were different in the two sides of the River Niger main axis. The endemic rate was low (4.44 % of the species), with one endemic species for each of the two sides of the Niger Delta. Twenty-one species were uncertain to occur in the study area, with a few being extinct (for instance, Choeropsis liberiensis). Co-occurrence models excluded that a competitively structured community of species is assembled in each of the vegetation zones, and cluster analyses suggested that biogeographic pressures may be important. 24.4 % of the extant species are globally threatened by IUCN, with one being Endangered (Pan troglodytes) and one Critically Endangered (Piliocolobus epieni). Flood forest and Eastern Flank forest represented the main vegetation zones inhabited by threatened species. In order to preserve the mammal community of the Delta, it is essential to carefully conserve flood and marsh forest patches, but also barrier island forests that still house highly threatened mammal populations.

Influence of exotic forest plantations on occupancy and co-occurrence patterns in a Mediterranean carnivore guild

The Mediterranean basin is a biodiversity hotspot which is being threatened by land abandonment and afforestation, most notably with eucalyptus (Eucalyptus spp.) plantations. We assessed the impact of eucalyptus plantations on niche partitioning in a carnivore community consisting of red fox (Vulpes vulpes), badger (Meles meles), and stone marten (Martes foina). Based on data from camera trapping, we evaluated the influence of eucalyptus plantations on species occupancy and detection in single-species and co-occurrence models and on temporal activity. Eucalyptus cover negatively influenced detection probability across all species in both single and co-occurrence models. Stone marten detection decreased with the presence of the other carnivores but red fox detection increased in the presence of badgers. Eucalyptus plantations had a negative effect on occupancy of red foxes, which preferred open farmland and evergreen oak forest. Stone marten preferred large patches of oak forest, whereas badger occupancy w...

CBAS: Context Based Arabic Stemmer

Arabic morphology encapsulates many valuable features such as word root. Arabic roots are being utilized for many tasks; the process of extracting a word root is referred to as stemming. Stemming is an essential part of most Natural Language Processing tasks, especially for derivative languages such as Arabic. However, stemming is faced with the problem of ambiguity, where two or more roots could be extracted from the same word. On the other hand, distributional semantics is a powerful co-occurrence model. It captures the meaning of a word based on its context. In this paper, a distributional semantics model utilizing Smoothed Pointwise Mutual Information (SPMI) is constructed to investigate its effectiveness on the stemming analysis task. It showed an accuracy of 81.5%, with a at least 9.4% improvement over other stemmers.