Predicting patients with Brain stroke
using the Logistic Regression classification Algorithm.

Data Source: Kaggle
Libraries and Packages: Scikit-learn, Python-Numpy, Pandas


Outline

Understanding the objective

Methodology

1. Data Setup

2. Data Cleaning

3. Defining a feature matrix

4. Building the model

5. Model Evaluation

6. Saving and Loading

7. Summary


Objective

Patients and medical attendants face a lot of uncertainity during, and after hospital encounters. Uncertainities like the possibilitty of a patient being diagnosed of an ailment after measuring certain vital signs. Machine learning models are capable of predicting a patient's outcome with up to a 90-95% accuracy. In this project, we will be building a model to predict and classify patients with the possibility of Brain stroke using the Logistic Regression Algorithm.

Features

Categorical features

Gender
Age
Marital status(Ever married)
Residence type
smoking status
Work type

Numerical features

Hypertension
heart_disease
avg_glucose_level
bmi

Prediction Variables

  • Age
  • Hypertension
  • heart disease
  • avg glucose level
  • bmi- body mass index

    • Output Variables

    The target shows the diagnosis of brain stroke: Displays whether the individual is suffering from brain stroke or not :
    0 = absence
    1 = present The score is being used as a conditional statement for classification as viewed in the deployed model section.

    1.Data Setup

    • Pip installing the dependencies.
    • Loading the dataset and required libraries

    2. Data Cleaning

    The data was inspected for null values to remove abnormality. However, the dataset showed no missing values.

    3. Defining a feature matrix

    Data was split into x and y to separate features to be trained from the response vector y The data was then split into training and testing.

    4. Building the model

    Check link The Logistic Regression was imported from scikitlearn and fit into the split data for training. Predict the Target variable

    5. Model Evaluation

    Model performance and Accuracy measurements was evalated using a number of performance metrics as shown below The model shows a classification accuracy score of 95%
    Comparing the Actual and Predicted response value for the first 30 predictions
    Confusion matrix
    Using the confusion matrix to calculate the classification metrics Printing a classification report
    The Area under the curve (AUC) score shows 85% accuracy. This is ususally adopted when the class is imbalanced.

    6. Saving and Loading the Model

    The trained model was saved and loaded with the pickle library for deployment.

    7. Summary of results

    A perfect model has an f1-score of 1.00. The model gave an f1-score of 0.98, perfect accuracy is equal to 1.0, we recorded an accuracy of 0.95 and AUC score of 85%. The actual and predicted value in comparison showed that the model is an excellent one for the classification is ready for deployment.

    Social