RMK ENGINEERING COLLEGE

An Autonomous Institution

Department of Computer Science and Engineering

Project Phase Review

MULTI DISEASE PREDICTION USING ML

Annapareddy Naveen (111720102008) Supervisor

Burle Nitesh (111720102028) Dr. D.Paulraj M.E, Ph.D.,
Chintam Sree Sasank Reddy (111720102029) Professor,
IV Year, Computer Science and Engineering Computer Science and Engineering
Date:27-01-2024
Content

1. Abstract

2. Scope

3. Literature Survey

4. Proposed system

5. Architecture

6. Modules description

7. References

2
 In healthcare, where existing machine learning models often focus on
single diseases, this study introduces a groundbreaking system that forecasts
multiple conditions, including diabetes, heart disease, and Parkinson's Disease.
The research utilizes diverse classification algorithms such as K-Nearest
Neighbor, Support Vector Machine, Decision Tree, Random Forest, Logistic
Regression, and Gaussian Naive Bayes. These algorithms enable precise
predictions and the early detection of potentially life-threatening conditions.
Multiple disease-specific datasets are employed to fine-tune the model's
ABSTRACT accuracy. The ultimate goal is to create a user-friendly web application that
consolidates disease predictions, offering a comprehensive healthcare solution.
This approach has the potential to revolutionize preventative medicine,
acknowledging the interconnected nature of health issues. However, ethical
concerns such as data privacy and algorithmic bias must be addressed. Overall,
this research signifies a significant advancement in healthcare and technology,
with the potential to save lives and improve public health.
 The multi-disease prediction system is a versatile and
user-friendly platform designed to anticipate a range of health
conditions by integrating data from diverse sources. With its
intuitive interface, real-time monitoring capabilities, and
adaptability to the dynamic landscape of healthcare, this system
provides a holistic approach to disease risk assessment. It caters to
the needs of both healthcare professionals and individuals, offering
SCOPE valuable insights into potential health issues. By harnessing data
from various origins, the system empowers users with the ability to
proactively manage their well-being and make informed decisions
about their health. In an era marked by evolving healthcare
demands, this system stands as a comprehensive tool, bridging the
gap between medical expertise and individual health awareness,
ultimately contributing to better health outcomes for all.
LITERATURE SURVEY
S.NO TITLE AUTHOR YEAR METHODOLOGY KEY POINTS INFERENCE

1 Heart Disease Jian Ping Li, 2020 This system has Heart Disease This paper suggests
Identification Method Amin Ul Haq, been designed for Classification, solutions to
Using Machine Learning Salah Ud Din, the identification Intelligent classifications problems
Classification in E- Jalaluddin Khan of heart disease. system. Using various
healthcare classification Algorithm

2 Heart Disease Prediction Vijeta Sharma, 2020 Only one module This System Heart Random
using Machine Manjari Gupta, Has been Achieves forest ,Heart support
LearningBayes Shrinkhala proposed in this 82.09 vector machine,Heart
Techniques Yadav work such as accuracy . Naïve Bayes prediction
Heart Disease models has been
proposed in this year

3 Multiple Heart Diseases Sateesh Ambes, 2021 Grid Search, This system Achievs
Prediction using Logistic Vijayalaxmi A, Turkey fence, 83.94 Accuracy
Regression with Ensemble Sridevi S, PCA,Feature
And Hyper Parameter Dr.Venkateswar Selection,
Tuning Techniques Logistic
Regression
S.NO TITLE AUTHOR YEAR METHODOLOGY KEY POINTS INFERENCE

4 Improving Heart Disease Saba Bashir, 2019 This system has Data Cleaning This paper suggests
Prediction Using Feature Zain Sikander, been designed for ,Binning,LR, solutions to
Selection Approaches Faran Hassan, Providing an Naïve Bayes. classifications problems
Aitzaz Anjum, accurate disease Using various
Khurram Bashir Diagnosis. classification Algorithm

5 Heart Disease Prediction P.Hamsagyathri 2020 Only one module This System This Paper Suggests
using SVM ,Sathyamangal- Has been Achieves solutions to Heart
am proposed in this 94.09 Disease using support
work such as accuracy . vector Machine
Heart Disease

6 An Effective Heart Disease Norma Latif, 2020 This paper review Heart disease, This system Achievs
Prediction Model for a Muhammad the impact of Outlier data, 82.87 Accuracy
Clinical Decision Support Syafrudin, Conventional Machine
System Ganjar Alfian. Prediction system learning.
 The proposed multi-disease prediction system integrates
vital patient data, employs advanced data integration techniques, and
utilizes machine learning algorithms including Naïve Bayes, Logistic
Regression, and Random Forest. It addresses the limitations observed
in existing systems by emphasizing data quality and consistency,
Proposed employing feature engineering, cross-validation, and regularization to
System optimize model performance, and hyper-tuning the features of the
data set. The system aims to provide a holistic, accurate, and practical
approach to multi-disease prediction, benefiting both healthcare
professionals and patients.
System
Architecture
 1.Heartdisease prediction

Modules 2.Diabetes prediction

3.Parkinson’s disease prediction
 The dataset we have gathered for our heart disease prediction project is
like a treasure chest full of valuable medical and clinical information. It's
designed with a primary focus on diagnosing and predicting heart diseases.
Inside this dataset, we find critical patient details such as age, gender,
blood pressure, cholesterol levels, and various indicators related to the
heart. It's a comprehensive compilation that offers a wealth of resources for
medical researchers, healthcare professionals, and data scientists. The
ultimate goal is to harness this information to create predictive models that
Module can assist in the early detection of heart diseases, thus potentially saving
Description lives and improving overall healthcare outcomes.
In this context, we have found that logistic regression is an excellent tool
Phase-I for binary classification, which is exactly what we need for our heart
(Heart-Disease) disease dataset. Binary classification is like a yes-or-no type of prediction,
and it's particularly well-suited for this dataset where we want to determine
whether a patient has heart disease or not. Logistic regression models the
likelihood of a patient having heart disease based on the information we
have about them. It's like having a smart assistant that can analyze all the
data and give us a probability of whether a patient is at risk or not. This
makes it a valuable asset for healthcare practitioners and researchers
because it provides a clear and interpretable way to make decisions about
patient care.
 Our ultimate goal in all of this is to improve patient care and healthcare
outcomes. Early detection of heart disease is crucial because it allows for
timely interventions and treatments, potentially saving lives and reducing
the burden on patients and healthcare systems. With a reliable predictive
model in place, healthcare practitioners can make more informed decisions,
Module and researchers can gather valuable insights. It's a win-win situation that
benefits both the medical community and patients.
Description To summarize, our dataset is a valuable resource filled with essential
Phase-I medical and clinical information. We're using this data to develop a
predictive model for heart disease detection, and we've chosen logistic
(Heart-Disease) regression as our tool of choice. This model is effective for binary
classification and can help us make predictions about a patient's heart
disease risk. By rigorously evaluating our model's performance using
various metrics, we ensure that it's reliable and accurate. Our efforts are all
geared towards improving patient care and healthcare outcomes, making a
positive impact on people's lives.
  Diabetes prediction
Module Our Diabetes Prediction Module utilizes machine learning algorithms
to assess diabetes risk, considering key metrics such as glucose
Description levels, BMI, blood pressure, age, gender, family history, physical
activity, cholesterol levels, and waist circumference. By synthesizing
Phase-II these factors, our model accurately predicts diabetes onset, enabling
(Diabetes timely interventions and personalized management strategies.
Through advanced ML algorithms like logistic regression and
prediction decision trees, healthcare professionals gain actionable insights for
early detection and intervention, enhancing patient outcomes and
) quality of life.
  Parkinson’s Disease:
Parkinson's Disease (PD) is a progressive neurodegenerative
disorder affecting millions worldwide. Early detection is paramount
Module for effective management. Our module integrates machine learning
(ML) with demographic and clinical information, including age,
Description gender, duration of symptoms, and family history of PD. Acoustic
features such as vocal fundamental frequency and jitter provide
Phase-III additional insights, aiding in accurate prediction and diagnosis.

(Parkinson’s By analyzing a combination of clinical parameters and acoustic

features extracted from voice recordings, our module offers a
Disease comprehensive approach to Parkinson's disease prediction. Smoking
status, medication history, and specific vocal characteristics like
) frequency variation and jitter play crucial roles in the predictive
model. This integration of ML with demographic and clinical
information empowers healthcare professionals with a robust tool for
early detection and personalized treatment strategies, ultimately
improving patient outcomes and quality of life.
Thank you