0% found this document useful (0 votes)
2 views4 pages

Heart Disease Prediction with Isolation Forest

The document outlines the author's internship projects, including the development of a User Login Anomaly Detection System and a Heart Disease Prediction Model using Isolation Forest and Random Forest algorithms. It highlights the use of ServiceNow for IT automation and Azure AutoML for model training. The author emphasizes the practical skills gained in AI/ML implementation, predictive modeling, and cloud automation during the internship.
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
2 views4 pages

Heart Disease Prediction with Isolation Forest

The document outlines the author's internship projects, including the development of a User Login Anomaly Detection System and a Heart Disease Prediction Model using Isolation Forest and Random Forest algorithms. It highlights the use of ServiceNow for IT automation and Azure AutoML for model training. The author emphasizes the practical skills gained in AI/ML implementation, predictive modeling, and cloud automation during the internship.
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

INTRODUCTION

During my NTCC project, I have successfully done all my assigned tasks in the
company,

I have:

 User Login Anomaly Detection System: Developed in an industry


environment using the Isolation Forest algorithm to detect abnormal login
behaviors such as unusual login times, multiple failed attempts, and
unknown devices. The model was trained on company-provided datasets,
and a Streamlit-based frontend was created to allow real-time input and
immediate anomaly detection, providing a practical security monitoring tool
for enterprise login systems.

 Heart Disease Prediction Model: Built and tested in the industry using the
Random Forest algorithm for predicting heart disease likelihood based on
medical parameters. A user-friendly frontend was developed to allow staff
and testing teams to input health parameters and receive instant predictions,
aiding in early awareness and corporate wellness initiatives.

 Explored and applied ServiceNow in the company, completing certifications


including AI Agent Fundamentals, Virtual Agent Basics, System
Administrator, and System

 Developer courses, earning badges and micro-certifications during the


internship period.
 Created and managed a ServiceNow personal developer instance within
the company environment, building and deploying a Virtual AI Agent for
Outlook Issues to automate troubleshooting, user guidance, and issue
escalation on the ServiceNow portal for internal IT support.

 Experimented with Azure AutoML under industry mentorship, learning


automated model training, testing, and deployment in live workflows to
support enterprise-level AI/ML projects.

These projects and certifications, completed during my company internship,


strengthened my practical, industry-aligned skills in advanced AI/ML
implementation, secure predictive modeling, frontend deployment, cloud
automation, and ServiceNow-based ITSM automation under real organizational
workflows.
Heart Disease Prediction Model

What is the project?

Heart Disease Prediction Model using Random Isolation Forest

 Objective: Predict the likelihood of heart disease using patient health parameters with an
anomaly detection approach.
 Instead of conventional classifiers (Random Forest, SVM), we use Isolation Forest
(iForest), typically used for anomaly detection, to identify outlier patterns associated
with heart disease.
 This method can be especially useful for early screening, flagging suspicious profiles
for further medical examination.

Why Isolation Forest for Heart Disease?

 Heart disease datasets often have imbalanced distributions (fewer disease cases).
 Isolation Forest:
o Efficient on high-dimensional data.
o No need for labeled anomalies during training.
o Works by isolating outliers, which are fewer and different (disease patterns)
from the majority (healthy).
o Faster and less memory-intensive compared to ensemble classifiers.
 Helps detect rare, hard-to-classify cases for early intervention.

Dataset Used

Typically uses UCI Heart Disease Dataset or Kaggle Heart Disease Dataset containing:

Feature Description
Age Age of the patient
Sex 1 = Male, 0 = Female
CP Chest Pain Type (0–3)
Trestbps Resting Blood Pressure
Chol Serum Cholesterol
Fbs Fasting Blood Sugar > 120 mg/dl
Restecg Resting ECG results
Thalach Maximum Heart Rate
Exang Exercise Induced Angina
Oldpeak ST depression induced by exercise
Feature Description
Slope Slope of the peak exercise ST segment
Ca Major vessels colored by fluoroscopy
Thal Thalassemia
Target 0 = No Heart Disease, 1 = Heart Disease

We remove Target during training to apply unsupervised anomaly detection, using it later for
evaluation.

Working Principle of Isolation Forest

 Randomly selects a feature and then randomly selects a split value between the
maximum and minimum values of the selected feature.
 Repeats recursively to isolate the data points.
 Points requiring fewer splits to isolate are considered anomalies (since they differ from
the majority).
 It computes an anomaly score between 0 and 1 for each point.
 Threshold tuning helps to convert anomaly scores into Yes/No predictions for heart
disease risk.

You might also like