Machine Learning for Heart Disease Prediction¶

Group 5:¶

Anke Li, Ce Guo, Samuel Merten, Yongseong Kim¶

Dataset: https://www.kaggle.com/datasets/bhaveshmisra/heart-disease-indicators/data¶

Questions:¶

  • How can we predict whether an individual has had an incidence of heart disease or a heart attack?
  • Which machine learning method demonstrates superior performance in detecting heart disease or heart attacks?

Methodology:¶

Given the many features within the dataset, there are a couple of initial steps that we’re considering taking to ensure robust predictions. For example, we’re thinking about using PCA to reduce dimensionality while still retaining relevant predictive features. Additionally, we may need to do some dataset balancing considering the low heart attack/disease incidence compared to non-incidence.

We are planning to implement the following machine learning models to the data:

  • K-Nearest Neighbors (KNN)
  • Support Vector Machine (SVM)
  • Decision Tree
  • Logistic Regression
  • Random Forest

We will assess the performance of each model using appropriate evaluation metrics, such as accuracy, precision, recall and ROC curve. Cross-validation techniques will be employed to ensure robustness in model assessment.

We expect to identify the machine learning algorithm(s) which perform best in predicting heart disease risk factors based on our evaluation metrics. The results will provide valuable insights into which model is most suitable for practical implementation in a healthcare setting.

Variable Descriptions:¶

  • HeartDiseaseorAttack: Indicates whether the individual has had a heart disease or heart attack (binary: 0 = No, 1 = Yes).
  • HighBP: High blood pressure status (binary: 0 = No, 1 = Yes).
  • HighChol: High cholesterol status (binary: 0 = No, 1 = Yes).
  • CholCheck: Frequency of cholesterol check (binary: 0 = No, 1 = Yes).
    • Have you had a cholesterol check within the last five years?
  • BMI: Body Mass Index (numeric).
  • Smoker: Smoking status (binary: 0 = No, 1 = Yes).
    • Have you smoked at least 100 cigarettes in your entire life?
  • Stroke: History of stroke (binary: 0 = No, 1 = Yes).
  • Diabetes: Diabetes status (binary: 0 = No, 1 = Yes).
  • PhysActivity: Level of physical activity (binary: 0 = No, 1 = Yes).
    • Have you engaged in physical activity or exercise during the past 30 days other than your regular job?
  • Fruits: Frequency of fruit consumption (binary: 0 = No, 1 = Yes).
    • Do you consume fruit 1+ times/day?
  • Veggies: Frequency of vegetable consumption (binary: 0 = No, 1 = Yes).
    • Do you consume veggies 1+ times/day?
  • HvyAlcoholConsump: Heavy alcohol consumption status (binary: 0 = No, 1 = Yes).
    • Determined as >=14 drinks/week for males, >=7 drinks/week for females.
  • AnyHealthcare: Access to any healthcare (binary: 0 = No, 1 = Yes).
  • NoDocbcCost: No doctor because of cost (binary: 0 = No, 1 = Yes).
    • Was there a time in the past 12 months when you needed to see a doctor but could not because of cost?
  • GenHlth: General health assessment (categorical).
    • Would you say that in general your health is: 1-5 rating. 1 = excellent, 5 = poor.
  • MentHlth: Mental health assessment (categorical).
    • Now thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good?
  • PhysHlth: Physical health assessment (categorical).
    • Now thinking about your physical health, which includes physical illness and injury, for how many days during the past 30 days was your physical health not good?
  • DiffWalk: Difficulty walking status (binary: 0 = No, 1 = Yes).
  • Sex: Gender of the individual (binary: 0 = Female, 1 = Male).
  • Age: Age of the individual (categorical).
    • 13 age categories ranging from 1: 18-24 years old to 13: 80+ years old
  • Education: Educational level (categorical).
    • 1 = Never attended school/only kindergarten
    • 2 = Grades 1-8
    • 3 = Grades 9-11
    • 4 = Grade 12 / GED
    • 5 = College 1-3 years
    • 6 = College 4+ years
  • Income: Income level (categorical).
    • Income scale with 8 categories ranging from 1: less than \$10,000 to 8: \\$75,000+

Reading in Data:¶

In [16]:
import pandas as pd

# Manually download dataset
heart = pd.read_csv("heart_disease_health_indicators.csv")

print(heart.info())

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 253661 entries, 0 to 253660
Data columns (total 22 columns):
 #   Column                Non-Null Count   Dtype
---  ------                --------------   -----
 0   HeartDiseaseorAttack  253661 non-null  int64
 1   HighBP                253661 non-null  int64
 2   HighChol              253661 non-null  int64
 3   CholCheck             253661 non-null  int64
 4   BMI                   253661 non-null  int64
 5   Smoker                253661 non-null  int64
 6   Stroke                253661 non-null  int64
 7   Diabetes              253661 non-null  int64
 8   PhysActivity          253661 non-null  int64
 9   Fruits                253661 non-null  int64
 10  Veggies               253661 non-null  int64
 11  HvyAlcoholConsump     253661 non-null  int64
 12  AnyHealthcare         253661 non-null  int64
 13  NoDocbcCost           253661 non-null  int64
 14  GenHlth               253661 non-null  int64
 15  MentHlth              253661 non-null  int64
 16  PhysHlth              253661 non-null  int64
 17  DiffWalk              253661 non-null  int64
 18  Sex                   253661 non-null  int64
 19  Age                   253661 non-null  int64
 20  Education             253661 non-null  int64
 21  Income                253661 non-null  int64
dtypes: int64(22)
memory usage: 42.6 MB
None
In [17]:
print(heart.describe())

       HeartDiseaseorAttack         HighBP       HighChol      CholCheck  \
count         253661.000000  253661.000000  253661.000000  253661.000000   
mean               0.094173       0.428990       0.424113       0.962667   
std                0.292070       0.494933       0.494209       0.189578   
min                0.000000       0.000000       0.000000       0.000000   
25%                0.000000       0.000000       0.000000       1.000000   
50%                0.000000       0.000000       0.000000       1.000000   
75%                0.000000       1.000000       1.000000       1.000000   
max                1.000000       1.000000       1.000000       1.000000   

                 BMI         Smoker         Stroke       Diabetes  \
count  253661.000000  253661.000000  253661.000000  253661.000000   
mean       28.382475       0.443186       0.040570       0.296904   
std         6.608638       0.496763       0.197292       0.698147   
min        12.000000       0.000000       0.000000       0.000000   
25%        24.000000       0.000000       0.000000       0.000000   
50%        27.000000       0.000000       0.000000       0.000000   
75%        31.000000       1.000000       0.000000       0.000000   
max        98.000000       1.000000       1.000000       2.000000   

        PhysActivity         Fruits  ...  AnyHealthcare    NoDocbcCost  \
count  253661.000000  253661.000000  ...  253661.000000  253661.000000   
mean        0.756577       0.634264  ...       0.951049       0.084164   
std         0.429149       0.481637  ...       0.215766       0.277633   
min         0.000000       0.000000  ...       0.000000       0.000000   
25%         1.000000       0.000000  ...       1.000000       0.000000   
50%         1.000000       1.000000  ...       1.000000       0.000000   
75%         1.000000       1.000000  ...       1.000000       0.000000   
max         1.000000       1.000000  ...       1.000000       1.000000   

             GenHlth       MentHlth       PhysHlth       DiffWalk  \
count  253661.000000  253661.000000  253661.000000  253661.000000   
mean        2.511379       3.184778       4.242028       0.168221   
std         1.068472       7.412822       8.717905       0.374063   
min         1.000000       0.000000       0.000000       0.000000   
25%         2.000000       0.000000       0.000000       0.000000   
50%         2.000000       0.000000       0.000000       0.000000   
75%         3.000000       2.000000       3.000000       0.000000   
max         5.000000      30.000000      30.000000       1.000000   

                 Sex            Age      Education         Income  
count  253661.000000  253661.000000  253661.000000  253661.000000  
mean        0.440348       8.032197       5.050461       6.054052  
std         0.496430       3.054203       0.985718       2.071036  
min         0.000000       1.000000       1.000000       1.000000  
25%         0.000000       6.000000       4.000000       5.000000  
50%         0.000000       8.000000       5.000000       7.000000  
75%         1.000000      10.000000       6.000000       8.000000  
max         1.000000      13.000000       6.000000       8.000000  

[8 rows x 22 columns]