My doctoral project addresses the theme of "Explainable Artificial Intelligence in Predicting Risk Factors Associated with Complications and Outcomes of Malaria". It is essential to explore the state of the art by reviewing related articles, understanding the approaches adopted by other researchers, learning about the models used, and understanding the types of data used, among other aspects.
Therefore, this repository was created to combine datasets used in malaria-related research, providing a valuable reference for future investigations. The focus will be exclusively on work utilizing machine learning on malaria-related biomedical datasets. Maintaining a data repository is a best practice and a fundamental component of data science and project development. This practice can boost transparency, foster collaboration, and increase efficiency, thus forming a solid foundation for future investigations and developments.
| abstract | Models used | Language | Year | Reference & Dataset |
|---|---|---|---|---|
| The article "Performance Analysis of Deep Learning Algorithms in Diagnosis of Malaria Disease" presents a comparative analysis of three deep learning models for diagnosing malaria. The models used were CNN, MobileNetV2, and ResNet50.The authors used a dataset of 27,558 images of blood cells, 13,780 images of parasitized cells, and 13,778 images of non-parasitized cells. The dataset was obtained from the United States National Institutes of Health (NIH) website.The analysis showed that the MobileNetV2 model performed best, with an accuracy rate of 97.06%. The CNN and ResNet50 models presented accuracy rates of 96.93% and 96.87%, respectively.The authors also used a data center to train the deep-learning models. | CNN (Convolutional Neural Network), MobileNetV2 e ResNet50 | English | 2023 | https://bmcmedimaging.biomedcentral.com/articles/10.1186/s12880-023-00993-9. The datasets used linkhttps://data.lhncbc.nlm.nih.gov/public/Malaria/Thick_Smears_150/index.html, is available here http://air.ug/downloads/plasmodium-phonecamera.zip, https://drive.google.com/drive/folders/1p45Dt-BJy8hhoI-rYnhcaL6IMl5FsFL-?usp=sharing (last accessed: 03/10/2022). https://data.lhncbc.nlm.nih.gov/public/Malaria/NIH-NLM-ThickBloodSmearsU/NIH-NLM-ThickBloodSmearsU.zip. This dataset was taken from the National Institutes of Health (NIH) website (https://www.nih.gov/, accessed on 11 August 2022). |
| Manual microscopic examination remains the gold standard for diagnosing malaria. However, this method is laborious and time-consuming and requires the expertise of trained pathologists. Deep learning (DL) has emerged as a promising alternative to manual microscopic examination for diagnosing malaria. In this study, the authors propose a tile-based deep-learning approach for malaria screening. The authors used a dataset of 11,290 thick blood smear images from Mali. The analysis results showed that the proposed approach achieves an accuracy of 96.83% for detecting and classifying malaria parasites, which is comparable to the accuracy of manual microscopic examination. | YOLOv4-S,YOLOv4-L | English | 2023 | https://bmcmedimaging.biomedcentral.com/articles/10.1186/s12880-023-00993-9. The datasets used linkhttps://data.lhncbc.nlm.nih.gov/public/Malaria/Thick_Smears_150/index.html, is available here http://air.ug/downloads/plasmodium-phonecamera.zip, https://drive.google.com/drive/folders/1p45Dt-BJy8hhoI-rYnhcaL6IMl5FsFL-?usp=sharing (last accessed: 03/10/2022). https://data.lhncbc.nlm.nih.gov/public/Malaria/NIH-NLM-ThickBloodSmearsU/NIH-NLM-ThickBloodSmearsU.zip. |
| https://data.lhncbc.nlm.nih.gov/public/Malaria/NIH-NLM-ThickBloodSmearsU/NIH-NLM-ThickBloodSmearsU.zip. | Decision Tree Classifier (DTC), Extra Tree Classifier (ETC), i.e. the article uses the hybrid microbe classifier (HMC): Combines the predictions of DTC, ETC. | English | 2023 | [Predicting microbe organisms using data of living micro forms of life and hybrid microbes classifier] dataset[ (https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0284522) https://doi.org/10.1371/journal.pone.0284522.s001 (ZIP)] |
| The study investigates the ability of machine learning to detect dengue virus infection (DENV) based solely on data available at the clinic visit. The data sets were used in the study "To Understand the Relationship with Work and the Occurrence of Malaria in Thailand Using Maximum Entropy Geospatial Modeling" to store and process the large amount of data collected for the study. The data included information on malaria cases, work locations, travel routes, and environmental factors. The data centers were also used to run the MaxEnt model, a machine-learning algorithm to predict the distribution of malaria cases in Thailand. | MaxEnt model | English | 2023 | (https://malariajournal.biomedcentral.com/articles/10.1186/s12936-023-04478-6) dataset (http://servir-rlcms.appspot.com/static/html/home.html), http://malaria.ddc.moph.go.th/malariar10/index_newversion.php |
| The primary objective of this study is to evaluate the ability of clinicians and machine learning algorithms to detect acute dengue virus (DENV) infection in febrile children in public clinics in Kenya. The study involved a large dataset of malaria cell images, which were collected from patients in malaria-endemic regions. The data center provided a secure and reliable environment to store and manage these images. The data center provided the ability to train and evaluate this CNN on a large dataset of malaria cell images. | Logistic Regression, Random Forest, Support Vector Machine (SVM), Gradient Boosting Machine (GBM), eXtreme Gradient Boosting (XGBoost), Artificial Neural Network (ANN) | English | 2023 | [Image analysis and machine learning-based malaria assessment system]](https://link.springer.com/article/10.1007/s12553-021-00620-z) dataset [(https://lhncbc.nlm.nih.gov/LHC-downloads/downloads.html#malariadatasets) |
| The article presents a system for diagnosing malaria using image analysis and machine learning. The system consists of three main components: image acquisition, image processing, and machine learning model. Machine learning model was: A convolutional neural network (CNN) was used to classify microscopic images of blood smears as positive or negative for malaria. The system achieved 90% accuracy on a dataset of 1,000 microscopic images of blood smears, which is significantly higher than the accuracy of traditional microscopy. A data center was used to train the machine learning algorithm to identify malaria parasites. And Evaluate the performance of the machine learning algorithm. The data center provided the ability to evaluate the performance of the machine learning algorithm. This involved testing the algorithm on a set of retained blood smear images to see how well it could identify malaria parasites. | Deep Convolutional Neural Network (CNN)Support Vector Machine(SVM) | English | 2023 | https://www.sciencedirect.com/science/article/pii/S2352864821000523. Data not available / Data will be made available on request |
| The main objective of the article "Predicting Plasmodium knowlesi transmission risk across Peninsular Malaysia using machine learning-based ecological niche modeling approaches" is to use machine learning-based ecological niche modeling techniques to predict the geographic distribution and transmission risk of the Plasmodium parasite knowlesi, the cause of simian malaria, throughout Peninsular Malaysia.The data center in this study was used for efficient storage, management, preprocessing, analysis, and visualization of large and complex data sets, which ultimately contributed to the development of predictive models for Plasmodium knowlesi transmission. risk in Peninsular Malaysia. | Maximum Entropy (MaxEnt), Random Forest (RF) | English | 2023 | https://www.frontiersin.org/articles/10.3389/fmicb.2023.1126418/full. Requests to access these datasets should be directed to [email protected]. The data analyzed in this study is subject to the following licenses/restrictions: The data for this study is available from the Ministry of Health Malaysia. Restrictions apply to the availability of this data. Data is available with permission from the Malaysian Ministry of Health. Data generated in the study are available from the corresponding author upon reasonable request. |
| The use of this data set was essential for the researchers to achieve their objectives in this study. The dataset allowed us to compare different algorithms, identify possible biases, and develop more robust models. The Article provides valuable insights into the use of deep learning for malaria diagnosis and may pave the way for the development of more accurate and reliable diagnostic tools.] | ResNet50, DenseNet121, InceptionV3 | English | 2023 | [Performance Analysis of Deep Learning Algorithms in Diagnosis of Malaria Disease (https://media.malariaworld.org/Performance_Analysis_of_Deep_Learning_Algorithms_in_Diagnosis_of_Malaria_Disease_758ec67bb1.pdf). This dataset was taken from the National Institutes of Health (NIH) website (https://www.nih.gov/, accessed on 11 August 2022). |
| The dataset included information on malaria cases in Thailand from 2019 to 2020. It was used to develop a geospatial malaria transmission model that was able to predict malaria transmission risk with high accuracy. | Maximum Entropy (MaxEnt) | English | 2023 | Understanding work-related travel and its relation to malaria occurrence in Thailand using geospatial maximum entropy modeling Landcover data is publicly available through http://servir-rlcms.appspot.com/static/html/home.html. Malaria cases data from MOPH is available through: http://malaria.ddc.moph.go.th/malariar10/index_newversion.php |
| The researchers in this paper used this dataset to train the XAI models, evaluate their performance, and generate explanations for their predictions. | Random Forest, Support Vector Machine (SVM) | Portuguese | 2023 | Predicting Plasmodium knowlesi transmission risk across Peninsular Malaysia using machine learning-based ecological niche modeling approaches dataset https://lhncbc.nlm.nih.gov/LHC-downloads/downloads.html#malariadatasets |
| This dataset was used to train and evaluate the performance of machine learning models. | Logistic Regression, Support Vector Machine (SVM), Random Forest | English | 2022 | [ Malaria cell image classification by explainable artificial intelligence ](https://www.tandfonline.com/doi/full/10.1080/08839514.2022.2031826) |
| The dataset used in this article was collected as secondary data from the Federal Polytechnic Ilaro Medical Center, Ilaro Ogun State, Nigeria, and contains information on 337 patients who presented for consultation regarding malaria-related infections. The symptoms reported by patients were.were recorded and information was collected on the same patients after they were tested for malaria.The recorded symptoms as reported by the patients were all compared with the malaria test results, and the malaria test results were used to target variables | Logistic Regression, Naive Bayes, SVM (Support Vector Machine), Decision Tree, Random Forest, K-nearest Neighbors Neural Network, AdaBoost. | English | 2022 | |
| The paper "Predicting Malaria Disease Using Frequency-Based Machine Learning Algorithms and Ensemble Algorithms" proposes a new approach to predicting malaria outbreaks using ensemble machine learning techniques. The article does not provide specific details about the data center used, but they do mention that the dataset used in the study was 10 years of malaria incidence data from Nigeria, which represents a large volume of data. | The paper proposes an ensemble model that combines four frequency-based machine-learning algorithms: Fast Fourier Transform (FFT), Discrete Wavelet Transform (DWT), Hilbert-Huang Transform (HHT), and Entropy-Based Frequency Analysis (EFA) | English | 2023 | |
| Dataset | Nr. Classes | Language | Year | Cite |
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent