# EC524: Heart-disease classification — Data

来源：https://www.kaggle.com/competitions/ec524-heart-disease/data

采集时间：2026-10-20T11:13:16Z

# Data

These data come from the [UCI Machine-Learning Repository](https://archive.ics.uci.edu/ml/index.php) (specifically from the [heart-disease dataset](https://archive.ics.uci.edu/ml/datasets/Heart+Disease).

The dataset that we will use has one outcome—`heart_disease`, an indicator for whether heart disease was present in the patient—and 13 features (predictors) on 303 patients.

## Outcome

The outcome `heart_disease` takes on either a `1` (heart disease is present) or `0` (not present). This variable is actually recoded from a variable that spanned 0 to 4, but we're going to focus on this recoding of 0/1.

## Features

The features (according to the original documentation)

1. `age`
2. `sex`
3. `chest_pain`: type of chest pain
- Value of 1: typical angina
- Value of 2: atypical angina
- Value of 3: non-anginal pain
- Value of 4: asymptomatic
4. `resting_bp`: resting blood pressure
5. `cholestoral`: serum cholestoral in mg/dl
6. `high_sugar`: an indicator for whether fasting blood sugar > 120 mg/dl  (1 = true; 0 = false)
7. `ecg`: resting electrocardiographic results
  - Value of 0: normal
  - Value of 1: having ST-T wave abnormality (T wave inversions and/or ST elevation or depression of > 0.05 mV)
  - Value of 2: showing probable or definite left ventricular hypertrophy by Estes' criteria
8. `max_rate`: maximum heart rate achieved
9. `exercise_angina`: exercise induced angina (1 = yes; 0 = no)
10. `st_depression`: ST depression induced by exercise relative to rest
11. `slope`: the slope of the peak exercise ST segment
  - Value of 1: upsloping
  - Value of 2: flat
  - Value of 3: downsloping
12. `vessels`: number of major vessels (0–3) colored by flourosopy
13. `thalium_scan`: thalium heart scan
  - Value of 3: normal
  - Value of 6: fixed defect
  - Value of 7: reversable defect

If you're wondering what *angina* is, [here is a description](https://www.heart.org/en/health-topics/heart-attack/angina-chest-pain) from the American Heart Association.

## Note: Life decisions

Clearly you can just download the dataset from its original repository and make perfect predictions. This strategy is cheating.