NTNorbert Tsopze

Norbert Tsopze

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Norbert Tsopze has a total of 28 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Multi-scale machine learning identifies indicator species from environmental-DNA co-occurrence networks

    by Eberrhard Estephe Kana Djifack, Eugenio Belda Cuesta, Thomas Lamy, Florian Batelaud, Norbert Tsopze, Jihad Zahir, Jean-Daniel Zucker, Laurent Vigliola, Edi Prifti

    Environmental DNA (eDNA) metabarcoding produces sparse read-count matrices routinely binarised to presence/absence to mitigate amplification, copy-number and shedding biases. We argue that binary eDNA data are sufficient to recover detailed community structure, sparse indicator panels and habitat-specific co-occurrence sub-networks — given the right analytical pipeline. Yet integrative frameworks bridging taxon-level and community-level structure remain missing for macro-organism eDNA. We propose a reproducible framework combining three scales: sparse machine-learning indicator-species identification (Predomics), consensus co-occurrence network inference (ScaleNet), and module-level enrichment via Gene Set Enrichment Analysis (GSEA). We applied it to 217 eDNA samples spanning eight habitats and three depth zones (shallow, <50 m; middle, 50–250 m; deep, 200–600 m) in New Caledonia, South Pacific. A sparse 55-MOTU indicator panel captured a proportion of beta-diversity variance comparable to the full 318-MOTU community. ScaleNet reconstructed a 261-node, 579-edge network with a core-periphery main component (shallow in the core, middle at the periphery) and a structurally isolated deep-zone subgraph. Fast-Greedy modularity partitioned it into 13 modules, 11 significantly enriched for depth zones or habitats by GSEA, each mapping onto a coherent trophic guild — from mesopelagic lanternfishes over deep seamounts to reef grazers across shallow microhabitats. Indicator taxa occupied structurally peripheral positions while connectivity hubs were non-indicator generalists, with reef-scraping parrotfish a functionally important exception. Sparse indicator panels efficiently capture zone discrimination while co-occurrence networks reveal community-level integration — together providing a scalable, interpretable framework for eDNA-based marine biodiversity monitoring.

  • Treillis de Galois et réseaux de neurones : une approche constructive d'architecture des réseaux de neurones

    by Norbert Tsopze

    Les réseaux de neurones artificiels connaissent des succès dans plusieurs domaines. Maisles utilisateurs des réseaux de neurones sont souvent confrontés aux problèmes de définitionde son architecture et d’interprétabilité de ses résultats. Plusieurs travaux ont essayé d’apporterune solution à ces problèmes. Pour les problèmes d’architecture, certains auteurs proposentde déduire cette architecture à partir d’un ensemble de connaissances décrivant le domaine duproblème et d’autres proposent d’ajouter de manière incrémentale les neurones à un réseauayant une taille initiale minimale. Les solutions proposées pour le problème d’interprétabilitédes résultats consistent à extraire un ensemble de règles décrivant le fonctionnement du réseau.Cette thèse contribue à la résolution de ces deux problèmes. Nous nous limitons à l’utilisationdes réseaux de neurones dans la résolution des problèmes de classification.Nous présentons dans cette thèse un état de l’art des méthodes existantes de recherche d’architecturede réseaux de neurones : une étude théorique et expérimentale est aussi faite. Decette étude, nous observons comme limites de ces méthodes la disponibilité absolue des connaissancespour construire un réseau interprétable et la construction des réseaux difficiles à interpréteren absence de connaissances. En alternative, nous proposons une méthode appelée CLANN(Concept Lattice-based Artificial Neural network) basée les treillis de Galois qui construit undemi-treillis à partir des données et déduire de ce demi-treillis l’architacture du réseau. CLANNétant limitée à la résolution des problèmes à deux classes, nous proposons MCLANN permettantd’étendre cette méthodes de recherche d’architecture des réseaux de neurones aux problèmes àplusieurs classes.Nous proposons aussi une méthode appelée ’Approche des MaxSubsets’ pour l’extractiondes règles à partir d’un réseau de neurones. La particularité de cette méthode est la possibilitéd’extraire les deux formats de règles (’si alors’ et ’m parmi N’) à partir d’une structure quenous construisons. Nous proposons aussi une façon d’expliquer le résultat calculé par le réseauconstruit par la méthode MCLANN au sujet d’un exemple.

  • Hybrid Method to Explain Predictions of Stacking Ensemble Model

    by Nguemkam Tebou, Ingrid Pamela, Norbert Tsopze, Dieudonné Tchuente


  • Epidemic threshold : a Laplacian spectral and structural approach of prediction

    by Claude Kanyou, Etienne Kouokam, Norbert Tsopze

    In epidemiology studies, control processes are driven by key parameters, such as \(R_0\), the epidemic threshold \(\tau\) over a contact network. By network-based models, the knowledge of network structures improves the prediction of \(\tau\), which is a challenge using structural features of a contact network. There are several structural approaches to predict \(\tau\). The common QMF (Quenched Mean-Field) approach uses the spectral radius as a single parameter. However, prediction can be improved using the node number, spectral radius, and Laplacian energy of graph. In this paper, at different levels, we design and experiment a new structural and spectral prediction approach of \(\tau\) called KSEL (K Spectral Energy of Laplacian). Theoretical and formal levels establish mathematical foundations, while qualitative, quantitative, and comparative levels compute a descriptive statistics summary, some data analytics, and visualisation through a large and heterogeneous dataset. Results show that the new approach effectively predicts \(\tau\). It captures the full network structure, connectivity, and network diffusion features. KSEL is similar, shares a common rolling trend, and performs really good compared to the previous structural prediction approaches, including the most commonly used QMF. There is a strong positive correlation and similar value distribution between KSEL and the previous structural prediction approaches that accepted the null hypothesis by ANOVA analysis. Therefore, the new approach is structurally enriched; it extends the structural and spectral area to analyse and control spreading processes over a network. The results can have practical interests to advise an effective epidemiological control policy.


  • Open Access

    Improving Tone Recognition Performance using Wav2vec 2.0-Based Learned Representation in Yoruba, a Low-Resourced Language

    by Saint Germes B. Bengono Obiang, Norbert Tsopze, Paulin Melatagia Yonta, Jean-Francois Bonastre, Tania Jiménez

    Many sub-Saharan African languages are categorized as tone languages, and for the most part, they are classified as low-resource languages due to the limited resources and tools available to process these languages. Identifying the tone associated with a syllable is therefore a key challenge for speech recognition in these languages. We propose models that automate the recognition of tones in continuous speech that can easily be incorporated into a speech recognition pipeline for these languages. We have investigated different neural architectures as well as several feature extraction algorithms in speech (FBs (Filter Banks), LEAF (Learnable Frontend), CS (Cestrogram), MFCC (Mel-Frequency Cepstral Coefficients)). In the context of low-resource languages, we also evaluated W2V (Wav2vec 2.0) models for this task. In this work, we use a public speech recognition dataset on Yoruba. As for the results, using the combination of features obtained from CS and FBs, we obtain a minimum TER (Tone Error Rate) of 19.54%, whereas for the evaluations of the models using W2V, we have a TER of 17.72%, demonstrating that the use of W2V provides better performance than the models used in the literature for tone identification on low-resource languages.


  • Explaining Meta-learner’s Predictions: Case of Corporate CO2 Emissions

    by Ingrid Pamela Nguemkam Tebou, Norbert Tsopze, Dieudonné Tchuente

  • UCAD: commUnity disCovery method in Attribute-based multicoloreD networks

    by Félicité Gamgne Domgue, Norbert Tsopze, René Ndoundam