CAPSTONE PROJECT

NONCOMMUNICABLE DISEASE PREDICTION

Presented By:
1.Manthra Sri D – SSMIET - CSE
2.Sherin Sithara M – SSMIET - CSE
3.Vaishnavi C – SSMIET - CSE
OUTLINE
◼ Problem Statement
◼ Proposed System/Solution
◼ System Development Approach
◼ Algorithm & Deployment
◼ Result
◼ Conclusion
◼ Future Scope
◼ References
PROBLEM STATEMENT

India has a high rate of Noncommunicable diseases (NCDs), which makes prompt and precise
monitoring of health indicators necessary. The majority of the time, the surveillance systems in use
today are insufficient, accurate, and thorough. When analyzing NCD indicators, employing machine
learning (ML) can improve predicted accuracy and offer more in-depth understanding of patterns and
risk factors. Creating strong machine learning models is crucial to enhancing public health initiatives,
timely intervention, and efficient NCD management in India.
1. Data Collection:

PROPOSED SOLUTION
 Gather historical data on NCDs, including patient demographics, medical history, lifestyle factors (e.g., smoking, diet, physical activity), and health
outcomes.
 Utilize real-time data sources, such as environmental conditions, air quality, social determinants of health, and healthcare access, to enhance prediction
accuracy.

2. Data Preprocessing:
 Clean and preprocess the collected data to handle missing values, outliers, and inconsistencies.

 Feature engineering to extract relevant features from the data that might influence the risk and prevalence of NCDs, such as socioeconomic factors,
comorbidities, and healthcare utilization patterns.

3. Machine Learning Algorithm:

 Implement a machine learning algorithm, such as a time-series forecasting model (e.g., ARIMA, SARIMA, or LSTM), to predict future NCD trends based
on historical patterns.
 Consider incorporating additional variables such as environmental factors, lifestyle choices, and demographic trends to improve the model’s predictive
accuracy.

4. Deployment:
 Develop a user-friendly interface or application that provides real-time predictions for NCD risk and prevalence, targeting different population groups or
regions.
 Deploy the solution on a scalable and reliable platform, considering factors like server infrastructure, data privacy regulations, and user accessibility.

5. Evaluation:
 Assess the model’s performance using appropriate metrics such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), or other relevant
evaluation criteria.
 Fine-tune the model based on feedback and continuous monitoring of prediction accuracy, adjusting for changes in population health trends and
external factors.
SYSTEM APPROACH
 Data Collection & Preprocessing: Gather health data from surveys, wearables, and healthcare records. Clean, normalize, and
perform feature engineering to handle missing values and standardize inputs.

 Risk Factor Analysis: Use algorithms to assess individual risk factors (BMI, age, blood pressure) for each disease and
calculate composite risk scores.

 Machine Learning Models: Implement classification models (e.g., logistic regression, random forests) to categorize risk levels
and regression models for predictions.

 Cloud Deployment: Deploy models on cloud platforms (AWS, Azure) for scalability, real-time data processing, and API
access.

 Data Visualization & Monitoring: Use dashboards for real-time tracking, trend analysis, and alerts for early intervention,
enabling authorities to make timely decisions on resource allocation and preventive measures.
ALGORITHM
1. Data Preprocessing
 Cleansing: Handle missing values, outliers, and inconsistencies.
 Feature Engineering: Create features like BMI, age, gender, family history, lifestyle factors (e.g., smoking, diet, exercise), and
environmental factors.
 Normalization: Scale continuous variables for standardization.

2. Risk Factor Calculation

 Risk scores will be computed for each disease using known risk factors:
 Cardiovascular Disease: Blood pressure, cholesterol, BMI, smoking, age, physical activity.
 Diabetes: BMI, age, family history, exercise, blood glucose.
 Cancer: Age, family history, diet, environmental exposure, smoking.
 These factors will be used to calculate a composite NCD risk score for each individual.

3. Machine Learning Models

 Classification: Logistic regression, decision trees, or random forests to categorize individuals into risk groups (low, medium, high).
 Regression: Linear regression or gradient boosting to predict continuous variables (e.g., cholesterol, blood pressure).
 Time Series: ARIMA or LSTM models to forecast future disease trends.

4. Indicator Tracking
 Track key NCD indicators (e.g., hypertension, obesity) over time, using historical and real-time data to predict emerging trends.
DEPLOYMENT
1. Cloud Infrastructure
Cloud Setup: Use AWS, Azure, or Google Cloud for data storage and processing.
Scalability: Cloud services like AWS Lambda or Google Cloud Functions will be used for real-time model predictions.
2. API Development
Develop RESTful APIs to provide NCD risk predictions.
Implement security protocols for data privacy.
3. Real-time Data Ingestion
Use data pipelines (e.g., Apache Kafka, AWS Kinesis) to collect real-time data from wearables, hospitals, and clinics.
4. Visualization and Reporting
Build dashboards with Dash or Power BI to visualize trends and predictions.
Enable filtering and analysis over different time periods.
5. Monitoring and Alerts
Set up alerts to notify health authorities when NCD indicators exceed thresholds.
Automate report generation for regular analysis.
RESULT

This project uses statistical and machine learning models to calculate NCD indicators based on factors
like BMI, age, lifestyle, and blood pressure. Data preprocessing ensures quality inputs, while risk scores
for diseases like cardiovascular conditions, diabetes, and cancer are computed for each individual.
Machine learning models predict NCD trends and classify risk levels, with real-time data ingestion and
visualization on a cloud-based platform. Health authorities can monitor indicators, analyze trends, and
receive alerts, enabling timely intervention and efficient resource allocation.
CONCLUSION

 This project will help build a robust, scalable system for monitoring non-communicable diseases
through health indicators. By leveraging advanced data science techniques and deploying the system
in a cloud-based environment, healthcare providers can make informed decisions about resource
allocation and interventions to combat NCDs. The system will also contribute to predictive health
analytics, which is crucial in managing the growing burden of chronic diseases worldwide.
FUTURE SCOPE

 To address the growing burden of non-communicable diseases (NCDs) in India, key indicators to
monitor include prevalence rates of conditions like diabetes, cardiovascular diseases, and cancer, along
with mortality and morbidity rates. Factors such as lifestyle, urbanization, and aging contribute to
rising NCD cases. Future strategies to cope should emphasize preventive measures, early screening,
health education, and accessible healthcare infrastructure, especially in rural areas. Strengthening
public health policies and promoting healthier lifestyles are also essential to control NCD-related risks
and reduce the economic and social impact of these diseases.
REFERENCES

 India faces a significant burden of non-communicable diseases (NCDs), with high rates of diabetes
(affecting around 101 million people), hypertension (315 million), and obesity, which impacts nearly
28.6% of adults. Additionally, about 24% of Indians suffer from hypercholesterolemia, increasing the
risk of cardiovascular diseases. Factors like urbanization, lifestyle changes, and limited access to
healthcare in rural areas contribute to the rise in NCDs. Addressing this requires better healthcare
infrastructure, especially in rural areas, improved access to early screening, and preventive health
policies​
THANK YOU