Brazilian Chikungunya Case Predictions with 26 Clinical Features, 2013-2020

12,709 notification records from Brazil's Information System for Notifiable Diseases (SINAN) between 2013 and 2020. The dataset includes 26 clinical symptom, comorbidity, and sociodemographic features for predicting laboratory-confirmed Chikungunya cases versus discarded suspected cases. It was created by Xinyu Lu and last updated in April 2026.

Use Cases

• Training diagnostic support models based on clinical symptom features like arthralgia and fever.
• Conducting SHAP-based interpretability analysis to identify dominant predictors of Chikungunya confirmation.
• Comparing model performance between algorithms like Random Forest and XGBoost for clinical deployment.
• Analyzing the impact of sociodemographic features such as age and education level on disease prediction.

Strengths

• Contains 12,709 nationally representative notification records from Brazil.
• Spans an 8-year time range from 2013 to 2020.
• Includes 26 distinct features covering clinical symptoms, comorbidities, and sociodemographics.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The model distinguishes confirmed Chikungunya from discarded suspected cases, not from other specific arboviruses.

Provenance

Source

Brazil's Information System for Notifiable Diseases (SINAN).

Collection Method

Derived from notification records.

Time Range

2013 to 2020

Freshness

Last updated 2026-04-13 04:30:40; freshness should be verified.

Geography

Brazil