# Colombian AI Olympiad 2025 (Pr1) Heart Attack — Data

来源：https://www.kaggle.com/competitions/colombian-ai-olympiad-2025-pr-1-heart-attack/data

采集时间：2026-10-20T07:26:10Z

**Dataset Description: "Heart Disease Risk Prediction Challenge"**

### **Contents**

The dataset contains **50,000 records** with **13 variables**, combining numerical, categorical, and temporal data:

| **Variable**         | **Type**        | **Description**                            | **Notes**                                  |
| -------------------- | --------------- | ------------------------------------------ | ------------------------------------------ |
| `age`                | Integer         | Patient’s age (20–80 years)                | Higher mean in class 1 (high risk)         |
| `gender`             | Binary          | Gender (0: Female, 1: Male)                | Higher proportion of males in class 1      |
| `smoker`             | Binary          | Is the patient a smoker? (0: No, 1: Yes)   | Probability conditional on gender and risk |
| `cholesterol`        | Continuous      | Cholesterol level (mg/dL)                  | Influenced by age and smoking status       |
| `cholesterol_change` | Continuous      | Year-to-year change in cholesterol (mg/dL) | Higher increase among smokers              |
| `stress_level`       | Ordinal         | Stress level ("low", "medium", "high")     | Encoded as 0, 1, 2                         |
| `diet_quality`       | Ordinal         | Diet quality ("poor", "average", "good")   | Encoded as 0, 1, 2                         |
| `heart_disease_risk` | Binary (Target) | Heart disease risk (0: Low, 1: High)       | Distribution: 40% class 1, 60% class 0     |

---

### **Key Dataset Features**

**Cascading dependencies:**

* Gender affects the probability of being a smoker.
* Smoking influences cholesterol levels and their temporal change.
* Age affects baseline cholesterol.

**Nonlinear variables:**

* `diet_quality` and `stress_level` are categorical with opposite distributions across classes.
* `cholesterol_change` depends on smoking status.

**Statistical realism:**

* Males have a higher risk, replicating real-world medical findings.
* Smokers in class 1 tend to have higher and increasing cholesterol levels.

We look forward to seeing **innovative solutions** that could help save lives in the real world!

---

This page accompanies the data files. It explains what files are provided and the format of each.
There’s no universal template for this section, but clear and detailed descriptions will help participants understand the data quickly and minimize clarification requests.

After reading this section, participants should be able to answer:

* What files do I need?
* What is the expected data format?
* What am I predicting?
* What do the columns and acronyms mean?

---

### **Files**

* **`train.csv`** – the training set
* **`test.csv`** – the test set
* **`sample_submission.csv`** – a sample submission file in the correct format