Mohammed Abebe Yimer
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Mohammed Abebe Yimer has a total of 16 co-authors. Here are the top co-authors by number of shared publications:
- Million Meshesha
School of Information Science, Addis Ababa University, Addis Ababa, Ethiopia
4 shared publications - Yohannes Minyilu
Faculty of Computing and Software Engineering, Institute of Technology, Arba Minch University, SE, Arba Minch, Ethiopia, 251 0911434681
4 shared publications - Abel Mekuria Molla
Dilla University
2 shared publications - Andualem Woldegiorgis
Faculty of Computing and Software Engineering, Arba Minch University, Arba Minch, Ethiopia
2 shared publications - Durga Prasad Sharma
Arba Minch University
2 shared publications - Suleyman Sevinc
Labenko Bilişim A.Ş.
2 shared publications
Latest Publications
- Open Access
- 22 October 2026
- Published by Public Library of Science (PLoS) in PLOS Mental Health
- Vol. 3 (10) , e0000721
A comparative analysis of machine learning, deep learning, and ensemble learning models for suicidal ideation detection on social media
Suicidal ideation is the consideration of self-inflicted death and has emerged as a serious and rising concern for worldwide public health systems, contributing to over 800,000 deaths annually worldwide. Individuals express intentions, distress, and thoughts of self-harm and death, thereby offering a crucial and unprecedented avenue for timely psychological intervention and surveillance. The core problem addressed by this research is the pronounced global insufficiency in detecting and mitigating suicidal ideation across linguistically diverse environments, specifically focusing on low-resource languages. Previous computational studies have primarily focused on English and other widely used languages, leaving a critical research gap for contexts such as Amharic. The main aim of this study was to investigate and to develop a robust and optimal deep learning model for identifying suicidal ideation in Amharic social media content. A total of 5,428 postings from social media such as Facebook and Telegram were carefully collected, manually annotated, and processed to provide the first publicly available dataset on Amharic suicidal ideation. After preprocessing, Word2Vec, GloVe, and FastText were among the word embedding techniques used for feature engineering. Five deep learning models and four machine learning models were then compared in this study. Based on the experimental results, ensemble learning models achieved slightly higher performance than the individual deep learning models in suicidal ideation detection. Specifically, hard voting ensemble learning models achieved the highest accuracy (83.3%) and precision (86.55%). CNN-BiLSTM achieved an accuracy of 81.45%, while the BiLSTM model achieved an accuracy of 81.4%, indicating that bidirectional sequence learning effectively captures contextual information in Amharic texts. Beyond its strong empirical performance, this study contributes a robust detection framework and introduces a public Amharic dataset, thereby establishing a foundational methodological resource for mental health research. Future studies should focus on developing scalable datasets and context-aware deep learning models that effectively handle multimodal data, multilingual transformer models, fusion-based features, and addressing bias in low-resource computational linguistics.
- 22 October 2026
- Published by Public Library of Science (PLoS) in PLOS Mental Health
- Vol. 3 (10) , e0000721
Open AccessA comparative analysis of machine learning, deep learning, and ensemble learning models for suicidal ideation detection on social media
Suicidal ideation is the consideration of self-inflicted death and has emerged as a serious and rising concern for worldwide public health systems, contributing to over 800,000 deaths annually worldwide. Individuals express intentions, distress, and thoughts of self-harm and death, thereby offering a crucial and unprecedented avenue for timely psychological intervention and surveillance. The core problem addressed by this research is the pronounced global insufficiency in detecting and mitigating suicidal ideation across linguistically diverse environments, specifically focusing on low-resource languages. Previous computational studies have primarily focused on English and other widely used languages, leaving a critical research gap for contexts such as Amharic. The main aim of this study was to investigate and to develop a robust and optimal deep learning model for identifying suicidal ideation in Amharic social media content. A total of 5,428 postings from social media such as Facebook and Telegram were carefully collected, manually annotated, and processed to provide the first publicly available dataset on Amharic suicidal ideation. After preprocessing, Word2Vec, GloVe, and FastText were among the word embedding techniques used for feature engineering. Five deep learning models and four machine learning models were then compared in this study. Based on the experimental results, ensemble learning models achieved slightly higher performance than the individual deep learning models in suicidal ideation detection. Specifically, hard voting ensemble learning models achieved the highest accuracy (83.3%) and precision (86.55%). CNN-BiLSTM achieved an accuracy of 81.45%, while the BiLSTM model achieved an accuracy of 81.4%, indicating that bidirectional sequence learning effectively captures contextual information in Amharic texts. Beyond its strong empirical performance, this study contributes a robust detection framework and introduces a public Amharic dataset, thereby establishing a foundational methodological resource for mental health research. Future studies should focus on developing scalable datasets and context-aware deep learning models that effectively handle multimodal data, multilingual transformer models, fusion-based features, and addressing bias in low-resource computational linguistics.
- Open Access
- 21 August 2026
- Published by JMIR Publications Inc. in Online Journal of Public Health Informatics
- Vol. 18, e84966
Automating Diagnosis of Skin Neglected Tropical Diseases via Patient Metadata through Machine Learning Model with Adaptive Balancing and Dual Cross-Validation: Retrospective Diagnostic Accuracy Study
Background: Skin neglected tropical diseases (NTDs) are the most prevalent diseases worldwide, affecting people living in resource-limited areas with low health care services and trained professionals. While machine learning (ML)–based diagnostic tools can be used for initial clinical assessment and patient screening, especially in resource-limited areas (including in Ethiopia), little effort has been made in this area. Objective: This pilot study develops a foundational ML model for the diagnosis of skin NTDs using patient metadata to analyze the feasibility of ML-based models for skin NTDs by identifying and experimentally evaluating 8 ML models. Methods: For this study, we acquired a tabular skin NTD diagnostic dataset collected from a specific affected district in the southwest of Ethiopia. We used the data in 3 different structures, which include using the initial dataset (IDS) that contains huge null values, using a final dataset (FDS) created through preliminary preprocessing, and a third dataset created by applying feature engineering (FEFDS). Selecting 8 ML models, we trained the models in 4 major experimental settings: baseline training, handling structural missing values, handling severe class imbalance through conditional class weighting, and a hybrid approach based on robust dual cross-validation (CV) consisting of an outer repeated stratified k -fold and nested CV methods. We used the macro and class-specific metrics (such as precision, recall, and F 1 -score), including balanced accuracy, due to the severe class imbalance. Feature importance scores are also used for evaluating overall model performance. Results: After the final training applying the hybrid approach, 4 models scored a perfect test score (1.0) across all the metrics and all experiments except naïve Bayes and multilayer perceptron (MLP), similarly scoring 0.997 balanced accuracy, 0.97 recall, and 0.985 F 1 -score. In the same experiment, the nested CV loop revealed a slight performance drop for light gradient boosting machine (LightGBM) and extreme gradient boosting (XGBoost), though both models similarly maintained higher mean scores of balanced accuracy and macro recall of 0.993 (SD 0.013), including the mean macro F 1 -score of 0.996 (SD 0.007), scoring declining performance with a mean score of 0.993 (SD 0.013) in balanced accuracy, macro recall, and F 1 -score, showing predictive biases. In terms of feature utilization, only CatBoost showed optimal feature ranking, while 4 models showed over-feature utilization (selection bias), with 3 models using small subsets of features (showing feature parsimony). Conclusions: Overall, this study has been highly challenged by data scarcity, class imbalances, limited disease representation, specific geographic representation, and lack of more data modalities. Hence, further studies are suggested to confirm the results on larger datasets having a representative distribution of disease classes and geographic locations.
- 21 August 2026
- Published by JMIR Publications Inc. in Online Journal of Public Health Informatics
- Vol. 18, e84966
Open AccessAutomating Diagnosis of Skin Neglected Tropical Diseases via Patient Metadata through Machine Learning Model with Adaptive Balancing and Dual Cross-Validation: Retrospective Diagnostic Accuracy Study
Background: Skin neglected tropical diseases (NTDs) are the most prevalent diseases worldwide, affecting people living in resource-limited areas with low health care services and trained professionals. While machine learning (ML)–based diagnostic tools can be used for initial clinical assessment and patient screening, especially in resource-limited areas (including in Ethiopia), little effort has been made in this area. Objective: This pilot study develops a foundational ML model for the diagnosis of skin NTDs using patient metadata to analyze the feasibility of ML-based models for skin NTDs by identifying and experimentally evaluating 8 ML models. Methods: For this study, we acquired a tabular skin NTD diagnostic dataset collected from a specific affected district in the southwest of Ethiopia. We used the data in 3 different structures, which include using the initial dataset (IDS) that contains huge null values, using a final dataset (FDS) created through preliminary preprocessing, and a third dataset created by applying feature engineering (FEFDS). Selecting 8 ML models, we trained the models in 4 major experimental settings: baseline training, handling structural missing values, handling severe class imbalance through conditional class weighting, and a hybrid approach based on robust dual cross-validation (CV) consisting of an outer repeated stratified k -fold and nested CV methods. We used the macro and class-specific metrics (such as precision, recall, and F 1 -score), including balanced accuracy, due to the severe class imbalance. Feature importance scores are also used for evaluating overall model performance. Results: After the final training applying the hybrid approach, 4 models scored a perfect test score (1.0) across all the metrics and all experiments except naïve Bayes and multilayer perceptron (MLP), similarly scoring 0.997 balanced accuracy, 0.97 recall, and 0.985 F 1 -score. In the same experiment, the nested CV loop revealed a slight performance drop for light gradient boosting machine (LightGBM) and extreme gradient boosting (XGBoost), though both models similarly maintained higher mean scores of balanced accuracy and macro recall of 0.993 (SD 0.013), including the mean macro F 1 -score of 0.996 (SD 0.007), scoring declining performance with a mean score of 0.993 (SD 0.013) in balanced accuracy, macro recall, and F 1 -score, showing predictive biases. In terms of feature utilization, only CatBoost showed optimal feature ranking, while 4 models showed over-feature utilization (selection bias), with 3 models using small subsets of features (showing feature parsimony). Conclusions: Overall, this study has been highly challenged by data scarcity, class imbalances, limited disease representation, specific geographic representation, and lack of more data modalities. Hence, further studies are suggested to confirm the results on larger datasets having a representative distribution of disease classes and geographic locations.
- Open Access
- 23 June 2026
- Published by JMIR Publications Inc. in JMIR Dermatology
- Vol. 9, e91544
Harmonized Dual Deep Learning Architectures for Image-Based Diagnostics of Skin Neglected Tropical Diseases: Benchmark Study via Novel Funnel Framework
Background: While deep learning–based methods are the potential technological solutions for the diagnosis of skin Neglected Tropical Diseases (skin NTDs), limited efforts were seen toward the use of such tools in Ethiopia. Data scarcity, methods, and models selection issues created further challenges in an attempt to close the previous gap. Objective: This study attempts to design a benchmark image-based diagnostic model for skin NTDs through a synergistic combination of feature extraction pretrained models, a custom-designed convolutional neural network (CNN) model trained on the extracted features, and an integrated data augmentation method applied dynamically. Methods: For this study, a new skin images dataset is created using skin photographs collected by a team of researchers from the NTDs research center of Arba Minch University Medical College. The new dataset contains 1495 images in 3 classes having severe class imbalance. Extensive experiments were conducted to find the optimal deep learning approach by designing a new CNN model, applying transfer learning, and designing the 2-stage approach that uses pretrained models for feature extraction and trains the new CNN model using the extracted features from the pretrained models and applying data augmentation based on the integrated 2-stage approach. For model selection, the study proposed a novel approach, the funnel framework with cascaded selection of methods and models. Results: After hyperparameter tuning, the model trained using DenseNet121 feature extractor scored the highest accuracy of 96.6%, F 1 -score of 95%, and sensitivity of 95%, while the MNv2-based model scored comparable results of 95.6% accuracy, 90% F 1 -score, and 90% sensitivity. This study finally selected the DenseNet121 and MNv2 models for feature extraction to build the final model for skin NTDs classification. Conclusions: The 2-stage approach significantly boosted the models’ performance compared with other methods, while the data augmentation method further enhanced the performance of the selected models. Finally, this study suggests further studies using advanced class-balancing methods with more data and a possible integration of other clinical data types.
- 23 June 2026
- Published by JMIR Publications Inc. in JMIR Dermatology
- Vol. 9, e91544
Open AccessHarmonized Dual Deep Learning Architectures for Image-Based Diagnostics of Skin Neglected Tropical Diseases: Benchmark Study via Novel Funnel Framework
Background: While deep learning–based methods are the potential technological solutions for the diagnosis of skin Neglected Tropical Diseases (skin NTDs), limited efforts were seen toward the use of such tools in Ethiopia. Data scarcity, methods, and models selection issues created further challenges in an attempt to close the previous gap. Objective: This study attempts to design a benchmark image-based diagnostic model for skin NTDs through a synergistic combination of feature extraction pretrained models, a custom-designed convolutional neural network (CNN) model trained on the extracted features, and an integrated data augmentation method applied dynamically. Methods: For this study, a new skin images dataset is created using skin photographs collected by a team of researchers from the NTDs research center of Arba Minch University Medical College. The new dataset contains 1495 images in 3 classes having severe class imbalance. Extensive experiments were conducted to find the optimal deep learning approach by designing a new CNN model, applying transfer learning, and designing the 2-stage approach that uses pretrained models for feature extraction and trains the new CNN model using the extracted features from the pretrained models and applying data augmentation based on the integrated 2-stage approach. For model selection, the study proposed a novel approach, the funnel framework with cascaded selection of methods and models. Results: After hyperparameter tuning, the model trained using DenseNet121 feature extractor scored the highest accuracy of 96.6%, F 1 -score of 95%, and sensitivity of 95%, while the MNv2-based model scored comparable results of 95.6% accuracy, 90% F 1 -score, and 90% sensitivity. This study finally selected the DenseNet121 and MNv2 models for feature extraction to build the final model for skin NTDs classification. Conclusions: The 2-stage approach significantly boosted the models’ performance compared with other methods, while the data augmentation method further enhanced the performance of the selected models. Finally, this study suggests further studies using advanced class-balancing methods with more data and a possible integration of other clinical data types.
- Open Access
- 16 June 2026
- Published by Springer Nature in Discover Artificial Intelligence
- Vol. 6 (1)
Developing an Amh3MD as a multimodal dataset for the detection of information disorders
- 16 June 2026
- Published by Springer Nature in Discover Artificial Intelligence
- Vol. 6 (1)
Open AccessDeveloping an Amh3MD as a multimodal dataset for the detection of information disorders
- Open Access
- 31 December 2025
- Published by Izmir Akademi Dernegi in Journal of Emerging Computer Technologies
- Vol. 5 (1) , 36-46
Customer Churn Prediction Using Machine Learning Techniques: Awash Bank Wolaita Sodo Region
Customer churn prediction refers to the procedure of identifying customers who are highly likely to terminate their service subscription based on their utilization. Being able to predict a customer who is likely to churn is essential for solving business problems. The banking industry in Ethiopia currently has millions of users, making it challenging to analyze and anticipate customer attrition. There are diverse researches conducted in this particular domain. The primary challenges encountered in the majority of the prior investigations were associated with the selection of suitable technique for achieving data balancing, the predicaments revolving around the choice of a technique for handling missing values, the excessive dependence of the model on a singular attribute, and various others. The aim of this research is to develop a machine-learning model that can predict customer churn. The dataset utilized for this investigation comprises 50,987 entries encompassing 11 attributes, which were collected from Awash Bank Wolaita Sodo region. Among these, 31,619 represent active accounts, while the remaining 19,368 pertain to closed (churn) accounts. To achieve balance within the dataset, a SMOTE-ENN method is employed, while an extraction tree classifier is employed for important feature selection. This research used an experimental research approach, and eight model are tested, including Extreme Gradient Boosting (XGBoost), random forest, Light Gradient-Boosting Machine (LightGBM), decision tree, Convolutional Neural Network (CNN), Gradient Boosting Machine (GBM), Deep Neural Network (DNN), and Multilayer Perceptron (MLP). Model performance is evaluated using accuracy, f1-score, recall, and precision. Experimental results show random forest model outperformed other models with an overall accuracy of 99.14% and recall, precision and f1-score of 99%.
- 31 December 2025
- Published by Izmir Akademi Dernegi in Journal of Emerging Computer Technologies
- Vol. 5 (1) , 36-46
Open AccessCustomer Churn Prediction Using Machine Learning Techniques: Awash Bank Wolaita Sodo Region
Customer churn prediction refers to the procedure of identifying customers who are highly likely to terminate their service subscription based on their utilization. Being able to predict a customer who is likely to churn is essential for solving business problems. The banking industry in Ethiopia currently has millions of users, making it challenging to analyze and anticipate customer attrition. There are diverse researches conducted in this particular domain. The primary challenges encountered in the majority of the prior investigations were associated with the selection of suitable technique for achieving data balancing, the predicaments revolving around the choice of a technique for handling missing values, the excessive dependence of the model on a singular attribute, and various others. The aim of this research is to develop a machine-learning model that can predict customer churn. The dataset utilized for this investigation comprises 50,987 entries encompassing 11 attributes, which were collected from Awash Bank Wolaita Sodo region. Among these, 31,619 represent active accounts, while the remaining 19,368 pertain to closed (churn) accounts. To achieve balance within the dataset, a SMOTE-ENN method is employed, while an extraction tree classifier is employed for important feature selection. This research used an experimental research approach, and eight model are tested, including Extreme Gradient Boosting (XGBoost), random forest, Light Gradient-Boosting Machine (LightGBM), decision tree, Convolutional Neural Network (CNN), Gradient Boosting Machine (GBM), Deep Neural Network (DNN), and Multilayer Perceptron (MLP). Model performance is evaluated using accuracy, f1-score, recall, and precision. Experimental results show random forest model outperformed other models with an overall accuracy of 99.14% and recall, precision and f1-score of 99%.
- Open Access
- 31 December 2025
- Published by Izmir Akademi Dernegi in Journal of Emerging Computer Technologies
- Vol. 5 (1) , 24-35
An Explainable Counterfeit and Genuine Ethiopian Banknote Classification
Counterfeiting is a serious crime with significant impact around the world and in Ethiopia in particular. The National Bank of Ethiopia has implemented various countermeasures to combat counterfeiting. The most successful counterfeit banknote detectors in use today are cash counters, which are hardware-based systems that use optical and magnetic sensors to detect and confirm banknotes. This technology entails an excessive cost and low availability for the public and small businesses, where the largest cash circulation occurs outside of banks. Existing countermeasures are insufficient to address this critical issue. Advancements in technology, such as digital printing and sophisticated scanning equipment, have made it easier for counterfeiters to mislead their victims by producing banknotes nearly identical to genuine ones. Only a handful of studies have been conducted on the current Ethiopian banknotes. This study presents an explainable deep learning-based model for the classification of genuine and counterfeit Ethiopian banknotes. The study used transfer learning with SHAP (Shapley Additive Explanations) and TF-EXPLAIN (TensorFlow Explain) explainable artificial intelligence frameworks for a better understanding of the classification prediction behind the models. Experimental results show that Dense121 achieved the best accuracy of 99.87%, and InceptionV3 achieved a remarkably similar result of 99.50%. To demonstrate the practical application of the model, a mobile application prototype was developed using Flutter and TensorFlow Lite. The application allows users to capture or upload images of banknotes for real-time classification without requiring internet connectivity. This solution provides an accessible and cost-effective counterfeit detection tool for the general public and small businesses.
- 31 December 2025
- Published by Izmir Akademi Dernegi in Journal of Emerging Computer Technologies
- Vol. 5 (1) , 24-35
Open AccessAn Explainable Counterfeit and Genuine Ethiopian Banknote Classification
Counterfeiting is a serious crime with significant impact around the world and in Ethiopia in particular. The National Bank of Ethiopia has implemented various countermeasures to combat counterfeiting. The most successful counterfeit banknote detectors in use today are cash counters, which are hardware-based systems that use optical and magnetic sensors to detect and confirm banknotes. This technology entails an excessive cost and low availability for the public and small businesses, where the largest cash circulation occurs outside of banks. Existing countermeasures are insufficient to address this critical issue. Advancements in technology, such as digital printing and sophisticated scanning equipment, have made it easier for counterfeiters to mislead their victims by producing banknotes nearly identical to genuine ones. Only a handful of studies have been conducted on the current Ethiopian banknotes. This study presents an explainable deep learning-based model for the classification of genuine and counterfeit Ethiopian banknotes. The study used transfer learning with SHAP (Shapley Additive Explanations) and TF-EXPLAIN (TensorFlow Explain) explainable artificial intelligence frameworks for a better understanding of the classification prediction behind the models. Experimental results show that Dense121 achieved the best accuracy of 99.87%, and InceptionV3 achieved a remarkably similar result of 99.50%. To demonstrate the practical application of the model, a mobile application prototype was developed using Flutter and TensorFlow Lite. The application allows users to capture or upload images of banknotes for real-time classification without requiring internet connectivity. This solution provides an accessible and cost-effective counterfeit detection tool for the general public and small businesses.
- Open Access
- 4 December 2025
- Published by JMIR Publications Inc. in JMIR AI
- Vol. 4, e67584
Application of Deep Learning-Based Multimodal Data Fusion for the Diagnosis of Skin Neglected Tropical Diseases: Systematic Review
Background: Neglected tropical diseases (NTDs) are the most prevalent diseases and comprise 21 different conditions. One-half of these conditions have skin manifestations, known as skin NTDs. The diagnosis of skin NTDs incorporates visual examination of patients, and deep learning (DL)–based diagnostic tools can be used to assist the diagnostic procedures. The use of advanced DL-based methods, including multimodal data fusion (MMDF) functionality, could be a potential approach to enhance the diagnostic procedures of these diseases. However, little has been done toward the application of such tools, as confirmed by the very few studies currently available that implemented MMDF for skin NTDs. Objective: This article presents a systematic review regarding the use of DL-based MMDF methods for the diagnosis of skin NTDs and related diseases (non-NTD skin diseases), including the ethical risks and potential risk of bias. Methods: The review was conducted based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) method using 6 parameters (research approach followed, disease[s] diagnosed, dataset[s] used, algorithm[s] applied, performance achieved, and future direction[s]). Results: Initially, 437 articles were collected from 5 major groups of identified sources; 14 articles were selected for the final analysis. Results revealed that, compared with traditional methods, the MMDF methods improved model performances for the diagnoses of skin NTDs and non-NTD skin diseases. Algorithmically, convolutional neural network (CNN)–based models were the predominantly used DL architectures (9/14 studies, 64% ), providing feature extraction, feature fusion, and disease classification, which were also conducted with transformer-based methods (1/14, 7%). Furthermore, recurrent neural networks were used in combination with CNN-based feature extractors to fuse multimodal data (1/14, 7%) and with generative models (1/14, 7%). The remaining studies used study-specific algorithms using transformers (1/14, 7%) and generative models (1/14, 7%). Conclusions: Finally, this article suggests that further studies should be conducted about using DL-based MMDF methods for skin NTDs, considering model efficiency, data scarcity, algorithm selection and use, fusion strategies of multiple modalities, and the possible adoption of such tools for resource-constrained areas.
- 4 December 2025
- Published by JMIR Publications Inc. in JMIR AI
- Vol. 4, e67584
Open AccessApplication of Deep Learning-Based Multimodal Data Fusion for the Diagnosis of Skin Neglected Tropical Diseases: Systematic Review
Background: Neglected tropical diseases (NTDs) are the most prevalent diseases and comprise 21 different conditions. One-half of these conditions have skin manifestations, known as skin NTDs. The diagnosis of skin NTDs incorporates visual examination of patients, and deep learning (DL)–based diagnostic tools can be used to assist the diagnostic procedures. The use of advanced DL-based methods, including multimodal data fusion (MMDF) functionality, could be a potential approach to enhance the diagnostic procedures of these diseases. However, little has been done toward the application of such tools, as confirmed by the very few studies currently available that implemented MMDF for skin NTDs. Objective: This article presents a systematic review regarding the use of DL-based MMDF methods for the diagnosis of skin NTDs and related diseases (non-NTD skin diseases), including the ethical risks and potential risk of bias. Methods: The review was conducted based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) method using 6 parameters (research approach followed, disease[s] diagnosed, dataset[s] used, algorithm[s] applied, performance achieved, and future direction[s]). Results: Initially, 437 articles were collected from 5 major groups of identified sources; 14 articles were selected for the final analysis. Results revealed that, compared with traditional methods, the MMDF methods improved model performances for the diagnoses of skin NTDs and non-NTD skin diseases. Algorithmically, convolutional neural network (CNN)–based models were the predominantly used DL architectures (9/14 studies, 64% ), providing feature extraction, feature fusion, and disease classification, which were also conducted with transformer-based methods (1/14, 7%). Furthermore, recurrent neural networks were used in combination with CNN-based feature extractors to fuse multimodal data (1/14, 7%) and with generative models (1/14, 7%). The remaining studies used study-specific algorithms using transformers (1/14, 7%) and generative models (1/14, 7%). Conclusions: Finally, this article suggests that further studies should be conducted about using DL-based MMDF methods for skin NTDs, considering model efficiency, data scarcity, algorithm selection and use, fusion strategies of multiple modalities, and the possible adoption of such tools for resource-constrained areas.
- Open Access
- 15 October 2024
- Published by JMIR Publications Inc. in JMIR Preprints
Application of Deep Learning-Based Multimodal Data Fusion for the Diagnosis of Skin Neglected Tropical Diseases: Systematic Review (Preprint)
BACKGROUND Neglected tropical diseases (NTDs) are the most prevalent diseases and comprise 21 different conditions. One-half of these conditions have skin manifestations, known as skin NTDs. The diagnosis of skin NTDs incorporates visual examination of patients, and deep learning (DL)–based diagnostic tools can be used to assist the diagnostic procedures. The use of advanced DL-based methods, including multimodal data fusion (MMDF) functionality, could be a potential approach to enhance the diagnostic procedures of these diseases. However, little has been done toward the application of such tools, as confirmed by the very few studies currently available that implemented MMDF for skin NTDs. OBJECTIVE This article presents a systematic review regarding the use of DL-based MMDF methods for the diagnosis of skin NTDs and related diseases (non-NTD skin diseases), including the ethical risks and potential risk of bias. METHODS The review was conducted based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) method using 6 parameters (research approach followed, disease[s] diagnosed, dataset[s] used, algorithm[s] applied, performance achieved, and future direction[s]). RESULTS Initially, 437 articles were collected from 5 major groups of identified sources; 14 articles were selected for the final analysis. Results revealed that, compared with traditional methods, the MMDF methods improved model performances for the diagnoses of skin NTDs and non-NTD skin diseases. Algorithmically, convolutional neural network (CNN)–based models were the predominantly used DL architectures (9/14 studies, 64% ), providing feature extraction, feature fusion, and disease classification, which were also conducted with transformer-based methods (1/14, 7%). Furthermore, recurrent neural networks were used in combination with CNN-based feature extractors to fuse multimodal data (1/14, 7%) and with generative models (1/14, 7%). The remaining studies used study-specific algorithms using transformers (1/14, 7%) and generative models (1/14, 7%). CONCLUSIONS Finally, this article suggests that further studies should be conducted about using DL-based MMDF methods for skin NTDs, considering model efficiency, data scarcity, algorithm selection and use, fusion strategies of multiple modalities, and the possible adoption of such tools for resource-constrained areas.
- 15 October 2024
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessApplication of Deep Learning-Based Multimodal Data Fusion for the Diagnosis of Skin Neglected Tropical Diseases: Systematic Review (Preprint)
BACKGROUND Neglected tropical diseases (NTDs) are the most prevalent diseases and comprise 21 different conditions. One-half of these conditions have skin manifestations, known as skin NTDs. The diagnosis of skin NTDs incorporates visual examination of patients, and deep learning (DL)–based diagnostic tools can be used to assist the diagnostic procedures. The use of advanced DL-based methods, including multimodal data fusion (MMDF) functionality, could be a potential approach to enhance the diagnostic procedures of these diseases. However, little has been done toward the application of such tools, as confirmed by the very few studies currently available that implemented MMDF for skin NTDs. OBJECTIVE This article presents a systematic review regarding the use of DL-based MMDF methods for the diagnosis of skin NTDs and related diseases (non-NTD skin diseases), including the ethical risks and potential risk of bias. METHODS The review was conducted based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) method using 6 parameters (research approach followed, disease[s] diagnosed, dataset[s] used, algorithm[s] applied, performance achieved, and future direction[s]). RESULTS Initially, 437 articles were collected from 5 major groups of identified sources; 14 articles were selected for the final analysis. Results revealed that, compared with traditional methods, the MMDF methods improved model performances for the diagnoses of skin NTDs and non-NTD skin diseases. Algorithmically, convolutional neural network (CNN)–based models were the predominantly used DL architectures (9/14 studies, 64% ), providing feature extraction, feature fusion, and disease classification, which were also conducted with transformer-based methods (1/14, 7%). Furthermore, recurrent neural networks were used in combination with CNN-based feature extractors to fuse multimodal data (1/14, 7%) and with generative models (1/14, 7%). The remaining studies used study-specific algorithms using transformers (1/14, 7%) and generative models (1/14, 7%). CONCLUSIONS Finally, this article suggests that further studies should be conducted about using DL-based MMDF methods for skin NTDs, considering model efficiency, data scarcity, algorithm selection and use, fusion strategies of multiple modalities, and the possible adoption of such tools for resource-constrained areas.
- 26 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 144-149
Hadiyyissa Automatic Speech Recognition Using Deep Learning Approach
Speech Recognition focuses on developing advanced technologies that can recognize and translate spoken language into text. Automatic Speech Recognition (ASR) is an Artificial Intelligence (AI) technology that captures human voice from a mi-crophone and analyzes it based on spoken words to comprehend natural language. In Ethiopia, ASR models have been developed for specific local languages such as Amharic, Tigrigna, Wolaytta, and Afaan Oromo, but there has been a lack of research on ASR models for the Hadiyyisa language, which is also spoken in Ethiopia. The research study designed and developed an ASR model for the Hadiyyisa language. The researchers collected data from native speakers of the Hadiyyisa language and used Python programming to design and develop the deep learning model for the Hadiyissa ASR model. Different methods and techniques were used in the study to achieve the research objectives and complete the entire research process. The proposed model utilized deep learning algorithms including RNN, LSTM, BiLSTM, and CNN, along with suitable feature extraction techniques such as Mel Frequency Cepstral Coefficients (MFCC). The accuracy of the proposed ASR model was evaluated through experimentation, which demonstrated that the BiLSTM model outperformed the other deep learning models with an overall accuracy of 90.95 % Based on the experimental evidence, the proposed model was found to be better suited for ASR for the Hadiyyisa language.
- 26 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 144-149
Hadiyyissa Automatic Speech Recognition Using Deep Learning Approach
Speech Recognition focuses on developing advanced technologies that can recognize and translate spoken language into text. Automatic Speech Recognition (ASR) is an Artificial Intelligence (AI) technology that captures human voice from a mi-crophone and analyzes it based on spoken words to comprehend natural language. In Ethiopia, ASR models have been developed for specific local languages such as Amharic, Tigrigna, Wolaytta, and Afaan Oromo, but there has been a lack of research on ASR models for the Hadiyyisa language, which is also spoken in Ethiopia. The research study designed and developed an ASR model for the Hadiyyisa language. The researchers collected data from native speakers of the Hadiyyisa language and used Python programming to design and develop the deep learning model for the Hadiyissa ASR model. Different methods and techniques were used in the study to achieve the research objectives and complete the entire research process. The proposed model utilized deep learning algorithms including RNN, LSTM, BiLSTM, and CNN, along with suitable feature extraction techniques such as Mel Frequency Cepstral Coefficients (MFCC). The accuracy of the proposed ASR model was evaluated through experimentation, which demonstrated that the BiLSTM model outperformed the other deep learning models with an overall accuracy of 90.95 % Based on the experimental evidence, the proposed model was found to be better suited for ASR for the Hadiyyisa language.
- Open Access
- 30 September 2023
- Published by Association for Computing Machinery (ACM) in Digital Threats: Research and Practice
- Vol. 4 (3) , 1-27
Single and Hybrid-Ensemble Learning-Based Phishing Website Detection: Examining Impacts of Varied Nature Datasets and Informative Feature Selection Technique
To tackle issues associated with phishing website attacks, the study conducted rigorous experiments on RF, GB, and CATB classifiers. Since each classifier was an ensemble learner on their own; we integrated them into stacking and majority vote ensemble architectures to create hybrid-ensemble learning. Due to ensemble learning methods being known for their high computational time costs, the study applied the UFS technique to address these concerns and obtained promising results. Since the scalability and performance consistency of the phishing website detection system across numerous datasets is critical to combating various variants of phishing website attacks, we used three distinct phishing website datasets (DS-1, DS-2, and DS-3) to train and test each ensemble learning method to identify the best-performed one in terms of accuracy and model computational time. Our experimental findings reveal that the CATB classifier demonstrated scalable, consistent, and superior accuracy across three distinct datasets (attained 97.9% accuracy in DS-1, 97.36% accuracy in DS-2, and 98.59% accuracy in DS-3). When it comes to model computational time, the RF classifier was discovered to be the fastest when applied to all datasets, while the CATB classifier was discovered to be the second quickest when applied to all datasets.
- 30 September 2023
- Published by Association for Computing Machinery (ACM) in Digital Threats: Research and Practice
- Vol. 4 (3) , 1-27
Open AccessSingle and Hybrid-Ensemble Learning-Based Phishing Website Detection: Examining Impacts of Varied Nature Datasets and Informative Feature Selection Technique
To tackle issues associated with phishing website attacks, the study conducted rigorous experiments on RF, GB, and CATB classifiers. Since each classifier was an ensemble learner on their own; we integrated them into stacking and majority vote ensemble architectures to create hybrid-ensemble learning. Due to ensemble learning methods being known for their high computational time costs, the study applied the UFS technique to address these concerns and obtained promising results. Since the scalability and performance consistency of the phishing website detection system across numerous datasets is critical to combating various variants of phishing website attacks, we used three distinct phishing website datasets (DS-1, DS-2, and DS-3) to train and test each ensemble learning method to identify the best-performed one in terms of accuracy and model computational time. Our experimental findings reveal that the CATB classifier demonstrated scalable, consistent, and superior accuracy across three distinct datasets (attained 97.9% accuracy in DS-1, 97.36% accuracy in DS-2, and 98.59% accuracy in DS-3). When it comes to model computational time, the RF classifier was discovered to be the fastest when applied to all datasets, while the CATB classifier was discovered to be the second quickest when applied to all datasets.