Predicting Heart Disease Through Speech Recognition

Project Title

Predicting Heart Disease Through Speech Recognition

Project Area of Specialization Computer ScienceProject Summary

1. Introduction

Atrial fibrillation (also called AFib or AF) is a quivering or irregular heartbeat (arrhythmia) that can lead to blood clots, stroke, heart failure and other heart-related complications. The worldwide prevalence of atrial fibrillation is 37,574 million cases (0.51% of worldwide population), increased also by 33% during the last 20 years.

   2.  Purpose

The purpose of this document to present and defend our Final Year Project idea on Prediction of heart disease (AFib) through different speech data streams by using distributed Machine-learning schemes and show the results on an Android App.  Local data streams will train on different edge nodes that are finally test and merge with master trained model. Training and testing phase of heart data streams and speech streams are performed after the fusion phase. The user in the form of 5-10 minutes speech gives the input to the app. After processing of input with the help of trained model, the app predict that the user has the AFib disease or not.

     3.  Scope

• Optimality of Machine learning Algorithms.
• Refactoring of the system.
• Minimum computational cost of the algorithms.
• Limited time complexity of the algorithm.

4. Product Perspective

This product will be a mobile application and will provide a user-friendly interface so the user can perform the functions easily. To use the application, the user will have to log in using the log in information After that, user can login to the system and access the functionality provided by the application.

Project Objectives

1. 1. 1. Problem Statement

The problem of	Not timely awareness of the people with the AFib disease. Untreated atrial fibrillation doubles the risk of heart-related deaths and is associated with a 5-fold increased risk for stroke. Many patients are unaware that AFib is a serious condition.
Affects	Heart patients.
the impact of which is	Worldwide prevalence of atrial fibrillation is 37,574 million cases (0.51% of worldwide population), increased also by 33% during the last 20 years.
a successful solution would be	An application that timely predict the Atrial fibrillation (AFib) disease in a person with the help of speech recognition. The prediction of AFib heart disease done through different speech data streams by using distributed Machine-learning schemes and show the results on an Android App. Local data stream will trained on different edge nodes that are finally test and merge with master trained model. Training and testing phase of heart data streams and speech streams are performed after the fusion phase. The user in the form of 5-10 minutes speech gives the input to the app. After processing of input with the help of trained model, the app predict that the user has the AFib disease or not.

1. 1. 1. Product Position Statement

For	User
Who	For the need of optimality in report and for the convenience of user to check the disease (AFib) and get the results.
Predicting Heart disease through speech Recognition	This is real time mobile application that predicts the AFib disease.
That	User will not go to the hospital for checking the disease and even no need for the doctor user just open the application and give the required information and then get the result of having disease or not. Even not to pay money for checking the disease.
Unlike	The current methods detect the AFib through ECG of Heart that’s a very expensive and time-consuming methods.
Our Product	Our mobile application will predict the AFib disease through user speech.

The problem of

 Affects

the impact of which is

a successful solution would be

For

Who

Predicting Heart disease through speech Recognition

That

Unlike

Our Product

Project Implementation Method

1. 1. Functional Requirements
      1. Interface

The app will have simple and user-friendly interface.

1. 1. 1. Login/Sign up

The user will be able to login in his/her account and new user sign up to a new account.

1. 1. 1. Input basic Information

The app will allow user to input their name, age, gender or basic information.

1. 1. 1. Built-in speech recorder

The app will have the built-in speech recorder for the user.

1. 1. 1. Speech Input

The user will be able to input his/her voice for prediction of AF disease.

1. 1. 1. Predict the disease

The app will be able to predict after the speech recognition that the user will have the AFib disease or not with the help of trained model.

1. 1. 1. Trained model

The app will have a Customized Neural Network (CNN) for the prediction of AFib disease.

1. 1. 1. Show Prediction results

The app will show the prediction results of AFib disease to the user.

1. 1. 1. User’s Feedback

The user will give feedback about the issue of the app.

1. 1. Other Non-Functional Requirements

1. 1. 1. Availability

Our application will be available 24/7.

1. 1. 1. Testability

The mobile application should be testable. A separate test environment should be set up where testers and the Quality Assurance engineers can test the application for bugs and/or incomplete or missed requirements.

1. 1. 1. Maintainability

The application should be maintainable. The application should be able to adapt new features. Previous functionality should not be disturbed in case of adding new functionality.

1. 1. 1. Supportability

Our mobile application will be supported on Android phone.

1. 1. 1. Accessibility

 The mobile application should be accessible to the users.

Benefits of the Project

An application that timely predict the Atrial fibrillation (AFib) disease in a person with the help of speech recognition. The prediction of AFib heart disease done through different speech data streams by using distributed Machine-learning schemes and show the results on an Android App. Local data stream will trained on different edge nodes that are finally test and merge with master trained model. Training and testing phase of heart data streams and speech streams are performed after the fusion phase. The user in the form of 5-10 minutes speech gives the input to the app. After processing of input with the help of trained model, the app predict that the user has the AFib disease or not.

Technical Details of Final Deliverable

There is a lot of disease that will go costly means there is huge checkup payment and even an ordinary man can’t pay that payment. Through is application a person can check their AFIB disease and get the result whether the person have this disease or not. This application will be used in hospitals and clinics. Through this application a patient will get a result in no time. A person checks their AFIB disease at home without going to hospital or went to doctor. So, this application provides a convenient solution to check AFIB disease.

An application that timely predict the Atrial fibrillation (AFib) disease in a person with the help of speech recognition. The prediction of AFib heart disease done through different speech data streams by using distributed Machine-learning schemes and show the results on an Android App. Local data stream will trained on different edge nodes that are finally test and merge with master trained model. Training and testing phase of heart data streams and speech streams are performed after the fusion phase. The user in the form of 5-10 minutes speech gives the input to the app. After processing of input with the help of trained model, the app predict that the user has the AFib disease or not.

Final Deliverable of the Project Software SystemCore Industry EducationOther Industries IT , Medical , Health Core Technology OthersOther TechnologiesSustainable Development Goals Good Health and Well-Being for PeopleRequired Resources

For	User
Who	For the need of optimality in report and for the convenience of user to check the disease (AFib) and get the results.
Predicting Heart disease through speech Recognition	This is real time mobile application that predicts the AFib disease.
That	User will not go to the hospital for checking the disease and even no need for the doctor user just open the application and give the required information and then get the result of having disease or not. Even not to pay money for checking the disease.
Unlike	The current methods detect the AFib through ECG of Heart that’s a very expensive and time-consuming methods.
Our Product	Our mobile application will predict the AFib disease through user speech.