Mushagalusa Ciza Arsène

Dr. Mushagalusa Ciza Arsène

orcid

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
28 June 2024 in Environmental Systems Research
Verified
Journal Article
Open access
PDF file
MAMushagalusa Ciza Arsène
Mushagalusa Ciza Arsène
AFAdandé Belarmain Fandohan
Adandé Belarmain Fandohan
Romain Lucas Glèlè KakaïRomain Lucas Glèlè Kakaï
Romain Lucas Glèlè Kakaï
Random forests (RF) have been widely used to predict spatial variables. Several studies have shown that spatial cross-validation (CV) methods consistently cause RF to yield larger prediction errors compared to standard CV methods. This study examined the impact of species characteristics and data features on the performance of the standard RF and spatial CV approaches for predicting species abundance distribution. It compared the standard 5-fold CV, design-based validation, and three different spatial CV methods, such as spatial buffering, environmental blocking, and spatial blocking. Validation samples were randomly selected for design-based validation without replacement. We evaluated their predictive performance (accuracy and discrimination metrics) using artificial species abundance data generated by a linear function of a constant term ( $$\beta _0$$ β 0 ) and a random error term following a zero-mean Gaussian process with a covariance matrix determined by an exponential correlation function. The model was tuned over multiple simulations to consider different mean levels of species abundance, spatial autocorrelation variation, and species detection probability. Here we found that the standard RF had poor predictive performance when spatial autocorrelation was high and the species probability of detection was low. Design-based validation and standard K-fold CV were found to be the most effective strategies for evaluating RF performance compared to spatial CV methods, even in the presence of high spatial autocorrelation and imperfect detection for random samples. For weakly or moderately clustered samples, they yielded good modelling efficiency but overestimated RF’s predictive power, while they overestimated modelling efficiency, predictive power, and accuracy for strongly clustered samples with high spatial autocorrelation. Globally, the checkerboard pattern in the allocation of blocks to folds in blocked spatial CV was found to be the most effective CV approach for clustered samples, whatever the degree of clustering, spatial autocorrelation, or species abundance class. The checkerboard pattern in spatial CV was found to be the best method for random or systematic samples with spatial autocorrelation, but less effective than non-spatial CV approaches. Failing to take data features into account when validating models can lead to unrealistic predictions of species abundance and related parameters and, therefore, incorrect interpretations of patterns and conclusions. Further research should explore the benefits of using blocked spatial K-fold CV with checkerboard assignment of blocks to folds for clustered samples with high spatial autocorrelation.

1 December 2022 in Journal of Probability and Statistics
Verified
Journal Article
Open access
PDF file
MAMushagalusa Ciza Arsène
Mushagalusa Ciza Arsène
AFAdandé Belarmain Fandohan
Adandé Belarmain Fandohan
Romain Lucas Glèlè KakaïRomain Lucas Glèlè Kakaï
Romain Lucas Glèlè Kakaï
Machine learning algorithms, especially random forests (RFs), have become an integrated part of the modern scientific methodology and represent an efficient alternative to conventional parametric algorithms. This study aimed to assess the influence of data features and overdispersion on RF regression performance. We assessed the effect of types of predictors (100, 75, 50, and 20 continuous, and 100 categorical), the number of predictors (p=816 and 24), and the sample size (N=50, 250, and 1250) on RF parameter settings. We also compared RF performance to that of classical generalized linear models (Poisson, negative binomial, and zero-inflated Poisson) and the linear model applied to log-transformed data. Two real datasets were analysed to demonstrate the usefulness of RF for overdispersed data modelling. Goodness-of-fit statistics such as root mean square error (RMSE) and biases were used to determine RF accuracy and validity. Results revealed that the number of variables to be randomly selected for each split, the proportion of samples to train the model, the minimal number of samples within each terminal node, and RF regression performance are not influenced by the sample size, number, and type of predictors. However, the ratio of observations to the number of predictors affects the stability of the best RF parameters. RF performs well for all types of covariates and different levels of dispersion. The magnitude of dispersion does not significantly influence RF predictive validity. In contrast, its predictive accuracy is significantly influenced by the magnitude of dispersion in the response variable, conditional on the explanatory variables. RF has performed almost as well as the models of the classical Poisson family in the presence of overdispersion. Given RFs advantages, it is an appropriate statistical alternative for counting data.

1 January 2021 in Journal of Veterinary Science
Verified
Journal Article
Open access
PDF file
PBPatrick Bisimwa
Patrick Bisimwa
MDMichel Dione
Michel Dione
BBBisimwa Basengere
Bisimwa Basengere
MAMushagalusa Ciza Arsène
Mushagalusa Ciza Arsène
LSLucilla Steinaa
Lucilla Steinaa
JOJuliette Ongus
Juliette Ongus
Bisimwa PN, et al. J Vet Sci. 2021 Apr;22:e35. https://doi.org/10.4142/jvs.2021.22.e35

11 January 2021 in Onderstepoort Journal of Veterinary Research
Verified
Journal Article
Open access
PDF file
MAMushagalusa Ciza Arsène
Mushagalusa Ciza Arsène
EEEric Marcel Charles Etter
Eric Marcel Charles Etter
Mary-Louise PenrithMary-Louise Penrith
Mary-Louise Penrith
The article reviews the outbreaks and distribution of African swine fever (ASF) in South Africa since the first probable outbreak that occurred in the Koedoesrand Ward in 1926. Retrospective data on the ASF outbreaks in South Africa were obtained from the World Organisation for Animal Health (OIE) disease database and the South African veterinary services annual reports in addition to published articles and online sources. South Africa has experienced many outbreaks that can be divided into 2 time periods: the period before the development of the OIE diseases database (1993) and the period after. More than 141 outbreaks of ASF were reported during the first period. Since the development of OIE disease database, 72 outbreaks directly involving 2968 cases, 2187 dead and 2358 killed pigs mainly in smallholder pig farms were reported. The median number of cases for a given ASF outbreak is 17, but in 50% of outbreaks no pigs were killed for prevention. The most important ASF outbreak was reported in April 2014 in the Greater Zeerust district (North West province) involving 326 cases and 1462 killed pigs. However, the outbreak with highest mortality involving 250 pigs was reported in 2016 (Free State province). According to phylogenetic analysis, nine p72 genotypes (I, III, IV, VII, VIII, XIX, XX, XXI and XXII) have been identified in South Africa. Season-wise, more outbreaks were recorded during summer. It was also observed that the OIE disease database could contain errors that would have been introduced through compiled forms at country level. Spatiotemporal studies on ASF outbreaks in South Africa are therefore required in order to assess statistically and quantitatively the clustering of outbreaks over space and time.

12 June 2024 in Modeling Earth Systems and Environment
Verified
Journal Article
Closed access
PDF file
MAMushagalusa Ciza Arsène
Mushagalusa Ciza Arsène
AFAdandé Belarmain Fandohan
Adandé Belarmain Fandohan
Romain Lucas Glèlè KakaïRomain Lucas Glèlè Kakaï
Romain Lucas Glèlè Kakaï

26 October 2019 in Veterinary World
Verified
Journal Article
Open access
MBMUTWEDU BWANA
MUTWEDU BWANA
BBBernard Kanyungu Buuma
Bernard Kanyungu Buuma
MAMushagalusa Ciza Arsène
Mushagalusa Ciza Arsène
NBNtagereka Patrick Bisimwa
Ntagereka Patrick Bisimwa
NCNadège CIREZI
Nadège CIREZI
RARodrigue Basengere Ayagirwe
Rodrigue Basengere Ayagirwe
Aim: This study was conducted to assess the prevalence of calf fetal wastage and its economic implications at ELAKAT slaughterhouse, Bukavu, Democratic Republic of Congo (DR Congo) to fill the research gap in relationship with this thematic. The study investigations took place from May to September 2018. Materials and Methods: A cross-sectional survey was carried at the slaughterhouse. For each visit, the number of cattle slaughtered, the number of pregnant cows slaughtered, and gestational ages (stage of pregnancy of the dam, and estimated by crown-rump length) of the fetuses were recorded. Results: Out of the 1035 cattle slaughtered during the study period, 970 were females. A total of 255 fetuses were recovered, representing a fetal recovery rate of 26.28%. The study established that one fetus was lost out of 4.5 cows slaughtered, and most of fetuses recovered (58.1%) were in the first trimester of gestation while 29.1% and 12.8% were, respectively, in the second and third trimester. Their age varied from 1.2 to 8.6 months, with body length ranging from 14 cm to 92 cm while their weight varied from 1.0 kg to 23.0 kg. The economic loss associated with the total cattle fetal wastages was estimated at Congo Democratic Francs (CDF) 29,906,400 ($15,787.5) with a monthly average of CDF 5,981,280 ($3,157.5). Conclusion: These results attested that slaughtering pregnant cows constitute a strong constraint on cattle industry development in DR Congo. Urgent measures, such as adequate enforcement of legislations on routine veterinary examinations at slaughterhouses as well as livestock owner's sensitization, are required to avoid selling pregnant cows during calving season.