Data Science Dojo
Copyright (c) 2016 - 2019
Level: Intermediate
Recommended Use: Classification Models
Domain: Automobiles
Echocardiogram Data Set
Will the patient survive for at least one year after a heart attack?
Public Domain, Link
This intermediate level data set has 132 rows and 12 columns. The data set provides data that could be used for classifying if patients will survive for at least one year after a heart attack. All the patients suffered heart attacks at some point in the past. Some are still alive and some are not. The survival and still-alive variables, when taken together, indicate whether a patient survived for at least one year following the heart attack.
According to the source: The problem addressed by past researchers was to predict from the other variables whether or not the patient will survive at least one year. The most difficult part of this problem is correctly predicting that the patient will NOT survive. (Part of the difficulty seems to be the size of the data set.)
This data set is recommended for learning and practicing your skills in exploratory data analysis, data visualization, dealing with missing values and classification modelling techniques. Feel free to explore the data set with multiple supervised and unsupervised learning techniques. The Following data dictionary gives more details on this data set:
Data Dictionary
| Column Position | Atrribute Name | Definition | Data Type | Example | % Null Ratios |
|---|---|---|---|---|---|
| 1 | Survival | The number of months patient survived (has survived, if patient is still alive). Because all the patients had their heart attacks at different times, it is possible that some patients have survived less than one year but they are still alive. Check the second variable to confirm this. Such patients cannot be used for the prediction task mentioned above. | Quantitative | 11, 57, 26 | 2 |
| 2 | Still-alive | A binary variable that depicts if the patient is still alive (0: dead at end of survival period, 1: still alive) | Quantitative | 0, 1 | 1 |
| 3 | Age-at-heart-attack | Age in years when heart attack occurred | Quantitative | 71, 57, 62 | 4 |
| 4 | Pericardial-effusion | Pericardial effusion is fluid around the heart (0: no fluid, 1: fluid) | Quantitative | 0, 1 | 1 |
| 5 | Fractional-shortening | A measure of contracility around the heart, lower numbers are increasingly abnormal | Quantitative | 0.23, 0.13, 0.45 | 6 |
| 6 | Epss | E-point septal separation, another measure of contractility. Larger numbers are increasingly abnormal | Quantitative | 6, 22, 12.062 | 11 |
| 7 | Lvdd | Left ventricular end-diastolic dimension. This is a measure of the size of the heart at end-diastole. Large hearts tend to be sick hearts | Quantitative | 4.26, 4.23, 5.39 | 8 |
| 8 | Wall-motion-score | A measure of how the segments of the left ventricle are moving | Quantitative | 14, 22.5, 27 | 3 |
| 9 | Wall-motion-index | Equals wall-motion-score divided by number of segments seen. Usually 12-13 segments are seen in an echocardiogram. (Preferable to use this variable INSTEAD of the wall motion score) | Quantitative | 1, 1.625, 2 | 1 |
| 10 | Mult | A derivate variable (suggested that it can be ignored) | Quantitative | 0.558, 1, 1.003 | 3 |
| 11 | Name | The Name of the patient (Has benn replaced with "name" entirely) | Qualitative | "name" | 0 |
| 12 | Group | Group (Has been considered meaningless and suggested to ignore) | Quantitative | 1, 2 | 17 |
| 13 | Alive-at-1 | It is a Boolean-valued variable derived from the first two attributes. (0: patient was either dead after 1 year or had been followed for less than 1 year, 1: patient was alive at 1 year) | Quantitative | 1, 0 | 44 |
Acknowledgement
This data set has been sourced from the Machine Learning Repository of University of California, Irvine Echocardiogram Data Set (UC Irvine). The UCI page mentions the following author for providing the data set: Steven Salzberg (salzberg@cs.jhu.edu)