# Global Dengue Outbreak Prediction Challenge — Data

来源：https://www.kaggle.com/competitions/global-dengue-outbreak-prediction-challenge/data

本地 Meta Kaggle 快照；不是本次实时页面。

# Dataset Description

The dataset provided for the **Global Dengue Outbreak Prediction Challenge** is a compilation of individual health records, designed to facilitate the development of machine learning models to predict dengue fever diagnosis. Each record contains a set of symptoms commonly associated with dengue and a binary label indicating whether the individual was diagnosed with the disease.

Below is the structure of the dataset along with a brief description of each field:

- `Name`: The individual's name. This is a nominal variable used for identification purposes.
- `Fever`: Binary indicator of the presence of fever (1 for present, 0 for absent).
- `Headache`: Binary indicator of the presence of headache (1 for present, 0 for absent).
- `JointPain`: Binary indicator of the presence of joint pain (1 for present, 0 for absent).
- `Bleeding`: Binary indicator of the presence of bleeding (1 for present, 0 for absent).
- `Dengue`: Binary label indicating whether the individual was diagnosed with dengue (1 for positive diagnosis, 0 for negative diagnosis).

Here is an example of how the data appears:

```
Name,Fever,Headache,JointPain,Bleeding,Dengue
Jeremy Long,0,0,1,1,1
Cassandra Glover,1,1,1,0,1
Aaron Adams MD,0,1,1,0,0
Elizabeth Thomas,0,1,0,1,1
Mathew Perez,0,0,1,0,1
Jordan Andrews,0,1,1,1,0
Mike Cruz,0,0,1,0,0
```

## Important Notes:

- The dataset is anonymized, with any personally identifiable information removed.
- The dataset contains both symptomatic cases and controls to help with the identification of patterns.
- Participants are encouraged to explore feature engineering opportunities to improve their model's predictive power.
- The dataset is balanced with respect to the target variable to prevent model bias toward a particular outcome.

Participants should use this dataset to train their models to predict the 'Dengue' variable. The quality of the predictions will be evaluated based on the evaluation metric detailed on the competition's evaluation page.