遇见数据集

electricsheepafrica/malaria-diagnosis-subsaharan-africa-2024

收藏
Hugging Face2026-03-27 更新2026-03-29 收录
官方服务:

资源简介:

--- dataset_info: features: - name: age dtype: float64 - name: temperature_celsius dtype: float64 - name: fever_duration_days dtype: float64 - name: parasite_density dtype: float64 - name: haemoglobin_gdl dtype: float64 - name: sex dtype: string - name: age_group dtype: string - name: symptom_headache dtype: string - name: symptom_chills dtype: string - name: symptom_vomiting dtype: string - name: symptom_diarrhoea dtype: string - name: symptom_joint_pain dtype: string - name: prior_antimalarial dtype: string - name: itn_use dtype: string - name: country dtype: string - name: location_type dtype: string - name: season dtype: string - name: rdt_result dtype: string - name: microscopy_result dtype: string - name: plasmodium_species dtype: string - name: treatment_outcome dtype: string - name: confirmed_malaria dtype: int64 splits: - name: train num_examples: 499 - name: test num_examples: 99 task_categories: - tabular-classification language: - en tags: - africa - nigeria - kenya - malaria-diagnostics - synthetic - machine-learning - electric-sheep-africa license: other --- # Malaria Clinical Diagnosis & RDT Outcomes Bundle — Teaser Dataset This is the **public teaser** of the Malaria Clinical Diagnosis & RDT Outcomes Bundle dataset bundle. It contains the full schema, documentation, and a **499-row sample**. **The complete bundle** — including the full dataset (30,000 rows), trained xgboost model (AUC-ROC: 1.000), and fully-executed notebook — is available on Gumroad: 👉 **[Get the full bundle on Gumroad](https://kossisoro.gumroad.com/l/malaria-dx)** --- ## Abstract This pack provides a research-grade, ML-ready dataset for malaria clinical diagnosis and RDT outcome prediction in Sub-Saharan Africa, with a focus on Nigeria and Kenya. The dataset comprises 30,000 individual-level records (12,000 real-base + 18,000 synthetic augmentation) across 22 features spanning demographics, clinical presentation (fever, symptoms, haemoglobin), diagnostic test results (RDT, microscopy), parasite characteristics, and treatment outcomes. Every distribution parameter is traceable to a verified data source: WHO Global Health Observatory API, DHS Program API, or peer-reviewed publications (all verified March 2026). Key verified statistics anchoring the dataset: Nigeria malaria incidence 294.25/1,000 [211–397] vs Kenya 74.17/1,000 [37–131] (WHO GHO 2024); Nigeria RDT test positivity 70.7% vs Kenya 43.0% (computed from WHO GHO 2024); Nigeria RDT prevalence in children <5: 39.6% vs Kenya 4.4% (DHS API). The pack includes a baseline XGBoost diagnostic classifier, ONNX export for edge deployment, inference wrapper, and full paper-style documentation. The Timber C99 compilation story makes this pack ideal for demonstrating embedded ML deployment on low-power diagnostic devices. --- ## Dataset Card | Attribute | Value | |---|---| | **Full dataset rows** | 30,000 (12,000 real + 18,000 synthetic) | | **Teaser rows** | 598 (this download) | | **Features** | 21 | | **Target** | `confirmed_malaria` | | **Geography** | Nigeria, Kenya | | **Model AUC-ROC** | 1.000 (on held-out test set, real data only) | --- ## Methodology Summary All synthetic distribution parameters are grounded in peer-reviewed sources. Features are sampled from specified distributions (truncated normal, lognormal, categorical, Poisson, etc.) with parameters extracted from published literature. Validation and test sets contain real data only for evaluation integrity. See the full README in the Gumroad bundle for complete methodology. --- ## Limitations - Geographic scope limited to Nigeria, Kenya - Synthetic data may not capture complex multivariate interactions - Not intended for direct production deployment without live data validation - See full README in the Gumroad bundle for comprehensive limitations --- ## Citation ```bibtex @dataset{esa_malaria_diagnosis_subsaharan_africa_2024_2026, author = {{Electric Sheep Africa}}, title = {Malaria Clinical Diagnosis & RDT Outcomes Bundle}, year = {2026}, version = {1.0.0}, publisher = {Gumroad}, } ``` --- *Electric Sheep Africa — Building Africa's AI data layer.*

数据集信息: 特征: - 名称:年龄(age),数据类型:float64 - 名称:摄氏体温(temperature_celsius),数据类型:float64 - 名称:发热持续天数(fever_duration_days),数据类型:float64 - 名称:寄生虫密度(parasite_density),数据类型:float64 - 名称:血红蛋白浓度(haemoglobin_gdl),数据类型:float64 - 名称:性别(sex),数据类型:字符串 - 名称:年龄组(age_group),数据类型:字符串 - 名称:头痛症状(symptom_headache),数据类型:字符串 - 名称:寒战症状(symptom_chills),数据类型:字符串 - 名称:呕吐症状(symptom_vomiting),数据类型:字符串 - 名称:腹泻症状(symptom_diarrhoea),数据类型:字符串 - 名称:关节痛症状(symptom_joint_pain),数据类型:字符串 - 名称:既往抗疟用药史(prior_antimalarial),数据类型:字符串 - 名称:驱虫蚊帐(insecticide-treated net, ITN)使用情况(itn_use),数据类型:字符串 - 名称:国家(country),数据类型:字符串 - 名称:居住地类型(location_type),数据类型:字符串 - 名称:季节(season),数据类型:字符串 - 名称:快速诊断试验(rapid diagnostic test, RDT)结果(rdt_result),数据类型:字符串 - 名称:显微镜检查结果(microscopy_result),数据类型:字符串 - 名称:疟原虫种类(plasmodium_species),数据类型:字符串 - 名称:治疗转归(treatment_outcome),数据类型:字符串 - 名称:确诊疟疾(confirmed_malaria),数据类型:int64 划分集: - 名称:训练集(train),样本量:499 - 名称:测试集(test),样本量:99 任务类别: - 表格分类(tabular-classification) 语言: - 英语 标签: - 非洲 - 尼日利亚 - 肯尼亚 - 疟疾诊断(malaria-diagnostics) - 合成数据(synthetic) - 机器学习(machine-learning) - 非洲电羊(electric-sheaf-africa) 许可证:其他 # 疟疾临床诊断与快速诊断试验结果数据集套件 — 预览数据集 本文件为**疟疾临床诊断与快速诊断试验结果数据集套件**的公开预览版,包含完整的数据集架构、说明文档与一份包含499条样本的子集。 **完整套件**包含全量数据集(30000条样本)、训练完成的XGBoost模型(受试者工作特征曲线下面积(AUC-ROC)为1.000)与完整可运行的Jupyter笔记本,可在Gumroad平台获取:👉 **[在Gumroad获取完整套件](https://kossisoro.gumroad.com/l/malaria-dx)** --- ## 摘要 本套件提供了一套符合研究级标准、可直接用于机器学习的数据集,用于撒哈拉以南非洲地区的疟疾临床诊断与快速诊断试验结果预测,研究重点覆盖尼日利亚与肯尼亚。数据集包含30000条个体级记录(其中12000条为真实原始数据,18000条为合成增强数据),涵盖22项特征,涉及人口统计学信息、临床表现(发热、相关症状、血红蛋白水平)、诊断检测结果(快速诊断试验、显微镜检查)、寄生虫特征与治疗转归。所有分布参数均可追溯至经验证的数据源:世界卫生组织全球卫生观察站(WHO Global Health Observatory, GHO)API、人口与健康调查项目(DHS Program)API,或同行评议文献(所有数据源均于2026年3月完成验证)。支撑本数据集的关键验证统计数据包括:尼日利亚疟疾发病率为294.25/1000 [211–397],肯尼亚为74.17/1000 [37–131](世界卫生组织全球卫生观察站2024年数据);尼日利亚快速诊断试验阳性率为70.7%,肯尼亚为43.0%(基于世界卫生组织全球卫生观察站2024年数据计算得出);尼日利亚5岁以下儿童快速诊断试验流行率为39.6%,肯尼亚为4.4%(人口与健康调查项目API数据)。本套件包含基线XGBoost诊断分类器、适用于边缘部署的ONNX导出模型、推理封装工具与完整的论文式说明文档。借助Timber C99编译方案,本套件非常适合演示如何在低功耗诊断设备上部署嵌入式机器学习模型。 --- ## 数据集卡片 | 属性 | 取值 | |---|---| | **全量数据集样本量** | 30000条(12000条真实数据 + 18000条合成数据) | | **预览版样本量** | 598条(本下载包包含) | | **特征数量** | 21项 | | **预测目标** | `confirmed_malaria`(确诊疟疾) | | **研究地域** | 尼日利亚、肯尼亚 | | **模型受试者工作特征曲线下面积(AUC-ROC)** | 1.000(仅基于真实数据的留出测试集计算) | --- ## 方法学概要 所有合成数据的分布参数均基于同行评议文献设定。特征从指定分布(截尾正态分布、对数正态分布、分类分布、泊松分布等)中采样,分布参数均来自已发表文献。为保证评估的严谨性,验证集与测试集仅包含真实数据。完整的方法学细节请参阅Gumroad套件中的完整README文档。 --- ## 局限性说明 - 研究地域仅覆盖尼日利亚与肯尼亚 - 合成数据可能无法捕捉复杂的多变量交互关系 - 未经实时数据验证的情况下,不得直接将本数据集用于生产部署 - 完整的局限性说明请参阅Gumroad套件中的README文档 --- ## 引用格式 bibtex @dataset{esa_malaria_diagnosis_subsaharan_africa_2024_2026, author = {{非洲电羊团队(Electric Sheep Africa)}}, title = {Malaria Clinical Diagnosis & RDT Outcomes Bundle}, year = {2026}, version = {1.0.0}, publisher = {Gumroad}, } --- *非洲电羊团队(Electric Sheep Africa)—— 构建非洲人工智能数据基础设施*

提供机构:
electricsheepafrica
二维码
社区交流群
二维码
科研交流群
商业服务