Performance evaluation of seven multi-label classification methods on real-world patent and publication datasets

Purpose:Many science,technology and innovation(STI)resources are attached with several different labels.To assign automatically the resulting labels to an interested instance,many approaches with good performance on the benchmark datasets have been proposed for multi-label classification task in the literature.Furthermore,several open-source tools implementing these approaches have also been developed.However,the characteristics of real-world multi-label patent and publication datasets are not completely in line with those of benchmark ones.Therefore,the main purpose of this paper is to evaluate comprehensively seven multi-label classification methods on real-world datasets.Research limitations:Three real-world datasets differ in the following aspects:statement,data quality,and purposes.Additionally,open-source tools designed for multi-label classification also have intrinsic differences in their approaches for data processing and feature selection,which in turn impacts the performance of a multi-label classification approach.In the near future,we will enhance experimental precision and reinforce the validity of conclusions by employing more rigorous control over variables through introducing expanded parameter settings.Practical implications:The observed Macro F1 and Micro F1 scores on real-world datasets typically fall short of those achieved on benchmark datasets,underscoring the complexity of real-world multi-label classification tasks.Approaches leveraging deep learning techniques offer promising solutions by accommodating the hierarchical relationships and interdependencies among labels.With ongoing enhancements in deep learning algorithms and large-scale models,it is expected that the efficacy of multi-label classification tasks will be significantly improved,reaching a level of practical utility in the foreseeable future.Originality/value:(1)Seven multi-label classification methods are comprehensively compared on three real-world datasets.(2)The TextCNN and TextRCNN models perform better on small-scale datasets with more complex hierarchical structure of labels and more balanced document-label distribution.(3)The MLkNN method works better on the larger-scale dataset with more unbalanced document-label distribution.

Pervasive Attentive Neural Network for Intelligent Image Classification Based on N-CDE’s

The utilization of visual attention enhances the performance of image classification tasks.Previous attentionbased models have demonstrated notable performance,but many of these models exhibit reduced accuracy when confronted with inter-class and intra-class similarities and differences.Neural-Controlled Differential Equations(N-CDE’s)and Neural Ordinary Differential Equations(NODE’s)are extensively utilized within this context.NCDE’s possesses the capacity to effectively illustrate both inter-class and intra-class similarities and differences with enhanced clarity.To this end,an attentive neural network has been proposed to generate attention maps,which uses two different types of N-CDE’s,one for adopting hidden layers and the other to generate attention values.Two distinct attention techniques are implemented including time-wise attention,also referred to as bottom N-CDE’s;and element-wise attention,called topN-CDE’s.Additionally,a trainingmethodology is proposed to guarantee that the training problem is sufficiently presented.Two classification tasks including fine-grained visual classification andmulti-label classification,are utilized to evaluate the proposedmodel.The proposedmethodology is employed on five publicly available datasets,including CUB-200-2011,ImageNet-1K,PASCAL VOC 2007,PASCAL VOC 2012,and MS COCO.The obtained visualizations have demonstrated that N-CDE’s are better appropriate for attention-based activities in comparison to conventional NODE’s.

Dietary polyphenols reduced the allergenicity ofβ-lactoglobulin via non-covalent interactions:a study on the structure-allergenicity relationship

Studies showed that complexation of polyphenols with milk allergens reduced their immunogenic potential.However,the relationship between structures of polyphenols and their hypoallergenic effects on milk allergens in association with physiological and conformational changes of the complexes remain unclear.In this study,polyphenols from eight botanical sources were extracted to prepare non-covalent complexes withβ-lactoglobulin(β-LG),a major allergen in milk.The dominant phenolic compounds bound toβ-LG with a diminished allergenicity were identified to investigate their respective role on the structural and allergenic properties ofβ-LG.Extracts from Vaccinium fruits and black soybeans were found to have great inhibitory effects on the IgE-and IgG-binding abilities ofβ-LG.Among the fourteen structure-related phenolic compounds,flavonoids and tannins with larger MWs and multi-hydroxyl substituents,notably rutin,EGCG,and ellagitannins were more potent to elicit changes on the conformational structures ofβ-LG to decrease the allergenicity of complexedβ-LG.Correlation analysis further demonstrated that a destabilized secondary structure and protein depolymerization caused by polyphenol-binding were closely related to the allergenicity property of formed complexes.This study provides insights into the understanding of structure-allergenicity relationship ofβ-LG-polyphenol interactions and would benefit the development of polyphenol-fortified matrices with hypoallergenic potential.

A Spectral Convolutional Neural Network Model Based on Adaptive Fick’s Law for Hyperspectral Image Classification

Hyperspectral image classification stands as a pivotal task within the field of remote sensing,yet achieving highprecision classification remains a significant challenge.In response to this challenge,a Spectral Convolutional Neural Network model based on Adaptive Fick’s Law Algorithm(AFLA-SCNN)is proposed.The Adaptive Fick’s Law Algorithm(AFLA)constitutes a novel metaheuristic algorithm introduced herein,encompassing three new strategies:Adaptive weight factor,Gaussian mutation,and probability update policy.With adaptive weight factor,the algorithmcan adjust theweights according to the change in the number of iterations to improve the performance of the algorithm.Gaussianmutation helps the algorithm avoid falling into local optimal solutions and improves the searchability of the algorithm.The probability update strategy helps to improve the exploitability and adaptability of the algorithm.Within the AFLA-SCNN model,AFLA is employed to optimize two hyperparameters in the SCNN model,namely,“numEpochs”and“miniBatchSize”,to attain their optimal values.AFLA’s performance is initially validated across 28 functions in 10D,30D,and 50D for CEC2013 and 29 functions in 10D,30D,and 50D for CEC2017.Experimental results indicate AFLA’s marked performance superiority over nine other prominent optimization algorithms.Subsequently,the AFLA-SCNN model was compared with the Spectral Convolutional Neural Network model based on Fick’s Law Algorithm(FLA-SCNN),Spectral Convolutional Neural Network model based on Harris Hawks Optimization(HHO-SCNN),Spectral Convolutional Neural Network model based onDifferential Evolution(DE-SCNN),SpectralConvolutionalNeuralNetwork(SCNN)model,and SupportVector Machines(SVM)model using the Indian Pines dataset and PaviaUniversity dataset.The experimental results show that the AFLA-SCNN model outperforms other models in terms of Accuracy,Precision,Recall,and F1-score on Indian Pines and Pavia University.Among them,the Accuracy of the AFLA-SCNN model on Indian Pines reached 99.875%,and the Accuracy on PaviaUniversity reached 98.022%.In conclusion,our proposed AFLA-SCNN model is deemed to significantly enhance the precision of hyperspectral image classification.

Emerging structures and dynamic mechanisms ofγ-secretase for Alzheimer’s disease

γ-Secretase,called“the proteasome of the membrane,”is a membrane-embedded protease complex that cleaves 150+peptide substrates with central roles in biology and medicine,including amyloid precursor protein and the Notch family of cell-surface receptors.Mutations inγ-secretase and amyloid precursor protein lead to early-onset familial Alzheimer’s disease.γ-Secretase has thus served as a critical drug target for treating familial Alzheimer’s disease and the more common late-onset Alzheimer’s disease as well.However,critical gaps remain in understanding the mechanisms of processive proteolysis of substrates,the effects of familial Alzheimer’s disease mutations,and allosteric modulation of substrate cleavage byγ-secretase.In this review,we focus on recent studies of structural dynamic mechanisms ofγ-secretase.Different mechanisms,including the“Fit-Stay-Trim,”“Sliding-Unwinding,”and“Tilting-Unwinding,”have been proposed for substrate proteolysis of amyloid precursor protein byγ-secretase based on all-atom molecular dynamics simulations.While an incorrect registry of the Notch1 substrate was identified in the cryo-electron microscopy structure of Notch1-boundγ-secretase,molecular dynamics simulations on a resolved model of Notch1-boundγ-secretase that was reconstructed using the amyloid precursor protein-boundγ-secretase as a template successfully capturedγ-secretase activation for proper cleavages of both wildtype and mutant Notch,being consistent with biochemical experimental findings.The approach could be potentially applied to decipher the processing mechanisms of various substrates byγ-secretase.In addition,controversy over the effects of familial Alzheimer’s disease mutations,particularly the issue of whether they stabilize or destabilizeγ-secretase-substrate complexes,is discussed.Finally,an outlook is provided for future studies ofγ-secretase,including pathways of substrate binding and product release,effects of modulators on familial Alzheimer’s disease mutations of theγ-secretase-substrate complexes.Comprehensive understanding of the functional mechanisms ofγ-secretase will greatly facilitate the rational design of effective drug molecules for treating familial Alzheimer’s disease and perhaps Alzheimer’s disease in general.

The Tension Cosmology, Largest Cosmic Structures and Explosions of Supernovae from SST

Here, using the Scale-Symmetric Theory (SST) we explain the cosmological tension and the origin of the largest cosmic structures. We show that a change in value of strong coupling constant for cold baryonic matter leads to the disagreement in the galaxy clustering amplitude, quantified by the parameter S8. Within the same model we described the Hubble tension. We described also the mechanism that transforms the gravitational collapse into an explosion—it concerns the dynamics of virtual fields that lead to dark energy. Our calculations concern the Type Ia supernovae and the core-collapse supernovae. We calculated the quantized masses of the progenitors of supernovae, emitted total energy during explosion, and we calculated how much of the released energy was transferred to neutrinos. Value of the speed of sound in the strongly interacting matter measured at the LHC confirms that presented here model is correct. Our calculations show that the Universe is cyclic.

Deep learning based classification of rock structure of tunnel face

The automated interpretation of rock structure can improve the efficiency,accuracy,and consistency of the geological risk assessment of tunnel face.Because of the high uncertainties in the geological images as a result of different regional rock types,as well as in-situ conditions(e.g.,temperature,humidity,and construction procedure),previous automated methods have limited performance in classification of rock structure of tunnel face during construction.This paper presents a framework for classifying multiple rock structures based on the geological images of tunnel face using convolutional neural networks(CNN),namely Inception-ResNet-V2(IRV2).A prototype recognition system is implemented to classify 5 types of rock structures including mosaic,granular,layered,block,and fragmentation structures.The proposed IRV2 network is trained by over 35,000 out of 42,400 images extracted from over 150 sections of tunnel faces and tested by the remaining 7400 images.Furthermore,different hyperparameters of the CNN model are introduced to optimize the most efficient algorithm parameter.Among all the discussed models,i.e.,ResNet-50,ResNet-101,and Inception-v4,Inception-ResNet-V2 exhibits the best performance in terms of various indicators,such as precision,recall,F-score,and testing time per image.Meanwhile,the model trained by a large database can obtain the object features more comprehensively,leading to higher accuracy.Compared with the original image classification method,the sub-image method is closer to the reality considering both the accuracy and the perspective of error divergence.The experimental results reveal that the proposed method is optimal and efficient for automated classification of rock structure using the geological images of the tunnel face.

Structure and Classification of Haemocytes in the Bivalve Mollusc Meretrix meretrix

Light and electron microscopic studies were carried out in order to characterize haemocytes in the bivalve mollusc Meretrix meretrix. According to nucleus and cytoplasm characters, four types of haemocytes were recognized： agranular haemocytes, lymphoid haemocyte, large granular and small granular haemocytes. Agranular hamocyte is the main cell type, accounting for 75%. It is agranular with rich organelles in cytoplasm, including mitochondria, golgi body and endoplasmic reticulum. Glycogen deposits were usually found in this cell type. The number of lymphoid haemocyte accounts for 1% - 2%. This cell type is agranular and shows a high ratio of nucleus to cytoplasm. A few organelles were found. High electrondense granules with diameters of 0.2 - 0.5μm and rich organelles were found in small granular haemocyte. The proportion of this cell type is about 15%. Rich granules of high electron-dense with diameters of 0.8- 2.4μm were found in large granular haemocyte. The proportion of this cell type is about 10%, and the quantity of organelles is the least.

Classification Criteria of Narrow/Wide Ice-ResistantConical Structures Based on Direct Measurements

Ice-induced structural vibration generally decreases with an increase in structural width at the waterline. Definitions of wide/narrow ice-resistant conical structures, according to ice-induced vibration, are directly related to structure width, sea ice parameters, and clearing modes of broken ice. This paper proposes three clearing modes for broken ice acting on conical structures: complete clearing, temporary ice pile up, and ice pile up. In this paper, sea ice clearing modes and the formation requirements of dynamic ice force are analyzed to explore criteria determining wide/narrow ice-resistant conical structures. According to the direct measurement data of typical prototype structures, quantitative criteria of the ratio of a cone width at waterline(D) to sea ice thickness(h) is proposed. If the ratio is less than 30(narrow conical structure), broken ice is completely cleared and a dynamic ice force is produced; however, if the ratio is larger than 50(wide conical structure), the front stacking of broken ice or dynamic ice force will not occur.

Tertiary structure-based protein classificationby virtual-bond-angles series

Structure-based protein classification can be based on the similarities in primary, second or tertiary structures of proteins. A method using virtual-bond-angles series that transformed the protein space configuration into a sequence was used for the classification of three-dimensional structures oi proteins. By transforming the main chains formed by C^a atoms of proteins into sequences, the series of virtual-bond-angles corresponding to the tertiary structure of the proteins were constructed. Then a distance-based hierarchical clustering method similar to Ward method was introduced to classify these virtual-bond-angles series of proteins. 200 files of protein structures were selected from Brookheaven protein data bank, and 11 clusters were classified.

A classification method of building structures based on multi-feature fusion of UAV remote sensing images

In order to improve the accuracy of building structure identification using remote sensing images,a building structure classification method based on multi-feature fusion of UAV remote sensing image is proposed in this paper.Three identification approaches of remote sensing images are integrated in this method:object-oriented,texture feature,and digital elevation based on DSM and DEM.So RGB threshold classification method is used to classify the identification results.The accuracy of building structure classification based on each feature and the multi-feature fusion are compared and analyzed.The results show that the building structure classification method is feasible and can accurately identify the structures in large-area remote sensing images.

Source Rock Classification and the Basic Structure of Coal and Kerogen

In accordance with the confusion on classification of source rocks, the authors raised a source rock classification for its enriched and dispersed organic matter types based on both Alpern’s idea and maceral genesis/composition. The determined rock type is roughly similar to palynofacies of Combaz , whereas it is "rock maceral facies (for coal viz. coal facies)" in strictly speaking. Therefore, it is necessary to use the organic ingredients classification proposed by the authors so that it can be used for both maceral analysis and environment research . This source rock classification not only shows sedimentology and diagenetic changes but also acquires organic matter type even if hydrocarbon potential derived from maceral’s geochemical parameters. So, it is considered as genetic classification. The "rock maceral facies" may be transformed to sedimentary organic facies , which is used as quantitative evaluation means if research being perfect.Now, there are many models in terms of structure either for coal or for kerogen. In our opinion, whatever coal or kerogen ought be polymer, then we follow Combaz’s thought and study structure of amorphous kerogens which are accordance with genetic mechanism showing biochemical and geochemical process perfectly. Here, we use the time of flight secondary ion mass spectrometry (TOFSIMS) to expand Combaz’s models from three to five. They are also models for coal.

Classification of regional agricultural structure:method and its application to China

Regional agriculture is the basis of regional sustainable development, so sustainable regional agricultural development is essential to the sustainable development of the whole society and becomes the focus of global research. The classification of regional agricultural structure is the basic work of regional agriculture study. This paper constructs index system （27 indices） of regional agricultural structure types with the three big indices： natural resources, developmental level of the agro-economy, and agro-ecological conditions. This paper also endows weight to every sub-classification index by means of AHP and obtains the comprehensive evaluation value of the three types of indices with arithmetic average weight approach. The regional agricultural structure can be classified into eight types in accordance with this evaluation results. The empirical study of China shows that the 31 provinces （municipalities and autonomous regions） are of different agriculture structural types. Finally, countermeasures of sustainsable agricultural development are put forward for the different agriculture structure features.

Phylogeny,character evolution,and classification of Selaginellaceae(lycophytes)

Selaginella is the largest and most taxonomically complex genus in lycophytes.The fact that over 750 species are currently treated in a single genus makes Selaginellales/Selaginellaceae unique in pteridophytes.Here we assembled a dataset of six existing and newly sampled plastid and nuclear loci with a total of 684 accessions(74%increase of the earlier largest sampling)representing ca.300 species to infer a new phylogeny.The evolution of 10 morphological characters is studied in the new phylogenetic context.Our major results include:(1)the nuclear and plastid phylogenies are congruent with each other and combined analysis well resolved and strongly supported the relationships of all but two major clades;(2)the Sinensis group is resolved as sister to S.subg.Pulviniella with strong support in two of the three analyses;(3)most morphological characters are highly homoplasious but some characters alone or combinations of characters well define the major clades in the family;and(4)an infrafamilial classification of Selaginellaceae is proposed and the currently defined Selaginella s.l.is split into seven subfamilies(corresponding to the current six subgenera t the Sinensis group)and 19 genera(the major diagnosable clades)with nine new species-poor genera.We support the conservation of Selaginella with a new type,S.flabellata,to minimize nomenclatural instability.We provide a key to subfamilies and genera,images illustrating their morphology,their morphological and geographical synopses,a list of constituent species,and necessary new combinations.This new classification will hopefully facilitate communication,promote further studies,and help conservation.

The chemical structures and biological activities of indole diterpenoids

Indole diterpenoids(IDTs)are an essential class of structurally diverse fungal secondary metabolites,that generally appear to be restricted to a limited number of fungi,such as Penicillium,Aspergillus,Claviceps,and Epichloe species,etc.These compounds share a typical core structure consisting of a cyclic diterpene skeleton of geranylgeranyl diphos-phate(GGPP)and an indole ring moiety derived from indole-3-glycerol phosphate(IGP).3-geranylgeranylindole(3-GGI)is the common precursor of all IDTs.On this basis,it is modified by cyclization,oxidation,and prenylation to generate a large class of compounds with complex structures.These compounds exhibit antibacterial,anti-insect,and ion channel inhibitory activities.We summarized 204 compounds of IDTs discovered from various fungi over the past 50 years,these compounds were reclassified,and their biological activities were summarized.This review will help to understand the structural diversity of IDTs and provide help for their physiological activities.

DNEF:A New Ensemble Framework Based on Deep Network Structure

Deep neural networks have achieved tremendous success in various fields,and the structure of these networks is a key factor in their success.In this paper,we focus on the research of ensemble learning based on deep network structure and propose a new deep network ensemble framework(DNEF).Unlike other ensemble learning models,DNEF is an ensemble learning architecture of network structures,with serial iteration between the hidden layers,while base classifiers are trained in parallel within these hidden layers.Specifically,DNEF uses randomly sampled data as input and implements serial iteration based on the weighting strategy between hidden layers.In the hidden layers,each node represents a base classifier,and multiple nodes generate training data for the next hidden layer according to the transfer strategy.The DNEF operates based on two strategies:(1)The weighting strategy calculates the training instance weights of the nodes according to their weaknesses in the previous layer.(2)The transfer strategy adaptively selects each node’s instances with weights as transfer instances and transfer weights,which are combined with the training data of nodes as input for the next hidden layer.These two strategies improve the accuracy and generalization of DNEF.This research integrates the ensemble of all nodes as the final output of DNEF.The experimental results reveal that the DNEF framework surpasses the traditional ensemble models and functions with high accuracy and innovative deep ensemble methods.

An Efficient 3D CNN Framework with Attention Mechanisms for Alzheimer’s Disease Classification

Neurodegeneration is the gradual deterioration and eventual death of brain cells,leading to progressive loss of structure and function of neurons in the brain and nervous system.Neurodegenerative disorders,such as Alzheimer’s,Huntington’s,Parkinson’s,amyotrophic lateral sclerosis,multiple system atrophy,and multiple sclerosis,are characterized by progressive deterioration of brain function,resulting in symptoms such as memory impairment,movement difficulties,and cognitive decline.Early diagnosis of these conditions is crucial to slowing down cell degeneration and reducing the severity of the diseases.Magnetic resonance imaging(MRI)is widely used by neurologists for diagnosing brain abnormalities.The majority of the research in this field focuses on processing the 2D images extracted from the 3D MRI volumetric scans for disease diagnosis.This might result in losing the volumetric information obtained from the whole brain MRI.To address this problem,a novel 3D-CNN architecture with an attention mechanism is proposed to classify whole-brain MRI images for Alzheimer’s disease(AD)detection.The 3D-CNN model uses channel and spatial attention mechanisms to extract relevant features and improve accuracy in identifying brain dysfunctions by focusing on specific regions of the brain.The pipeline takes pre-processed MRI volumetric scans as input,and the 3D-CNN model leverages both channel and spatial attention mechanisms to extract precise feature representations of the input MRI volume for accurate classification.The present study utilizes the publicly available Alzheimer’s disease Neuroimaging Initiative(ADNI)dataset,which has three image classes:Mild Cognitive Impairment(MCI),Cognitive Normal(CN),and AD affected.The proposed approach achieves an overall accuracy of 79%when classifying three classes and an average accuracy of 87%when identifying AD and the other two classes.The findings reveal that 3D-CNN models with an attention mechanism exhibit significantly higher classification performance compared to other models,highlighting the potential of deep learning algorithms to aid in the early detection and prediction of AD.

Parkinson’s Disease Classification Using Random Forest Kerb Feature Selection

Parkinson’s disease(PD)is a neurodegenerative disease cause by a deficiency of dopamine.Investigators have identified the voice as the underlying symptom of PD.Advanced vocal disorder studies provide adequate treatment and support for accurate PD detection.Machine learning(ML)models have recently helped to solve problems in the classification of chronic diseases.This work aims to analyze the effect of selecting features on ML efficiency on a voice-based PD detection system.It includes PD classification models of Random forest,decision Tree,neural network,logistic regression and support vector machine.The feature selection is made by RF mean-decrease in accuracy and mean-decrease in Gini techniques.Random forest kerb feature selection(RFKFS)selects only 17 features from 754 attributes.The proposed technique uses validation metrics to assess the performance of ML models.The results of the RF model with feature selection performed well among all other models with high accuracy score of 96.56%and a precision of 88.02%,a sensitivity of 98.26%,a specificity of 96.06%.The respective validation score has an Non polynomial vector(NPV)of 99.47%,a Geometric Mean(GM)of 97.15%,a Youden’s index(YI)of 94.32%,and a Matthews’s correlation method(MCC)90.84%.The proposed model is also more robust than other models.It was also realised that using the RFKFS approach in the PD results in an effective and high-performing medical classifier.

Temporal change of plankton size structure preserved by Lugol's solution:a FlowCAM study

Plankton size structure is crucial for understanding marine ecosystem dynamics and the associated biogeochemical processes.A fixation step by acid Lugol’s solution has been commonly employed to preserve plankton samples in the field.However,the acid Lugol’s solution can bias the estimation of size structure and the preserved plankton size structure can vary with time.Here,we explore the impact of sample storage time on the size-structure of the plankton community preserved by Lugol’s solution.Two short-term experiments and one long-term experiment were conducted to explore the change of plankton community size structure with the storage time:covering from a week to a month,and to nearly seven months based on particle-size data obtained by continuous Flow Cytometer and Microscope(FlowCAM)measurements.We found a linear change of plankton size with the storage time in short-term periods(less than 3 months)with a decrease of the slope but an increase of the intercept for the normalized biomass size spectrum(NBS S).However,there were opposite trends for NBSS with increasing slope but decreasing intercept after3 months.The potential causes of the distinct patterns of the NBSS parameters are addressed in terms of the interplay between particle aggregation and fragmentation.We found large changes in plankton biovolume and abundance among different size classes,which may indicate a distinct effect of acid Lugol’s solution on various plankton size classes.The mechanism driving temporal change in the size-structure of the Lugolfixed plankton community was further discussed in terms of particle aggregation and fragmentation.Finally,we emphasize that the effect of storage time should be taken into account when interpreting or comparing data of plankton community acquired from samples with various storage durations.

Cesarean Sections according to the Robson’s Classification in Two University Hospitals of Yaoundé: Indications and Maternofetal Outcome

Introduction: Cesarean section is a surgical intervention which consists in the extraction of a fetus from the uterus after its incision. The rate of cesarean section varies depending on the country and the health facility. For this reason, in 2015, the World Health Organization (WHO) recommended the use of Robson’s classification to evaluate the practice of cesarean sections in order to identify the groups of women who had abnormally high rates. The objective of our study was to evaluate cesarean sections using the Robson’s classification in CHRACERH and in the Yaoundé Central Hospital (YCH). Methodology: We carried out a retrospective cross sectional and descriptive study in two (02) university hospitals in Yaoundé which took place from December 2017 to May 2018. We included in our study all women who gave birth over a period of two (02) years from January 2016 to December 2017 in these two health facilities. Our sampling was exhaustive over the study period. The parturients’ information was collected using an anonymous and pretested questionnaire. The Robson’s group of every parturient was determined. Descriptive parameters like mean and proportions were calculated. We compared the rates and indications of cesarean sections between the both hospitals using Chi2 test. Results: Out of 330 deliveries realized in CHRACERH, we had 90 cesarean sections;hence, a rate of 27.2%. Out of 1863 deliveries carried out at the YCH, 462 were by cesarean section, hence a rate of 24.8%. The women who belonged to groups 1, 3 and 5 contributed to the highest rates of cesarean sections in both hospitals: in CHRACERH, group 5 (31.1%), group 3 (20%) and group 1 (15.6%), at YCH: group 3 (22.5%), group 1 (21.6%) and group 5 (17.3%). The indications of the cesarean sections varied depending on the Robson’s group and the hospital, the principal indication in group 1 was acute fetal distress (28.6%) in CHRACERH and cephalopelvic disproportion (36.7%) at YCH. Cephalopelvic disproportion was the predominant indication in groups 3 of CHRACERH (44.4%) and YCH (39.2%). In groups 5, CHRACERH and of YCH, a scarred uterus was the principal indication for the cesarean section at 82.4% and 78.4% respectively. At CHRACERH, the maternofetal complications were more frequent in groups 1 and 2 at the YCH, this was the case mostly in groups 1 and 3. Conclusion: The Robson’s classification is an adequate tool for the evaluation and comparison of the rates of cesarean sections. The rates of cesarean section in CHRACERH (27.2%) and at YCH (24.8%) were higher than the rates recommended by WHO. Robson’s groups 1, 3 and 5 were identified as the groups most at risk for cesarean sections in the both hospitals.