Machine Learning-Based

Personalized Health Monitoring and

Predictive Analytics
 Md. Ashik Karim Nayon
ID: 0272310005101063

Sara Shahrin Moumi Rabia Akter Rutmila Meherin

ID: 0272310005101062 ID: 1915002047 ID: 2015302018
CONTENT
 Introduction
 Motivation
 Objectives
 Literature Review
 Problem Statement
 Methodology
 Conclusion
 Future Work
 Reference

1
INTRODUCTION
• Machine learning and predictive analytics have
emerged as powerful tools in the field of
healthcare, ushering in a new era of disease
prevention and transforming traditional
approaches to disease identification, diagnosis,
and treatment.
• These cutting-edge technologies possess the
potential to revolutionize the landscape of
healthcare, paving the way for more effective
preventive measures that can save countless
lives.
2
MOTIVATION
• To utilize machine learning classifiers on
symptom data in order to accurately predict
disease.
• To decide on the dataset’s most important
features in order to enhance prediction
performance.
• To provide individualized nutritional
recommendations for the successful
management of every illness.
• To provide personalized exercise plans
appropriate for various medical issues.
• To validate the system’s effectiveness in multi-
class disease classification and advice.

3
OBJECTIVES
• Emerging Opportunities in Healthcare

• Opportunity to improve patient

outcomes and reduce healthcare costs

• Contribution to Intelligent Healthcare

Systems

4
LITERATURE REVIEW
Paper Name Author Name Algorithm used Accuracy
Logistic Regression, Support Vector Machine,
Healthcare Predictive Analytics Using Machine
Mohammed Badawy , Nagy Ramadan and Random Forest, K-means Clustering, Deep
Learning and Deep Learning Techniques: A 87.72%
Hesham Ahmed Hefny Learning, Long Short-Term Memory, Hybrid
Survey
Approaches
Support Vector Machine (SVM), Random Forest
Shahadat Uddin, Arif Khan, Md Ekramul (RF), Naïve Bayes (NB), Artificial Neural
Comparing Different Supervised Machine
Hossain, Mohammad Ali Moni Network (ANN), Logistic Regression (LR), 53%
Learning Algorithms for Disease Prediction
Decision Tree (DT), and K-Nearest Neighbour
(KNN).
Random Forest, Support Vector Machines,
Seema S. Awathare, Samiksha G. Gajbhiye,
Heart Disease Prediction Using Machine Neural Network
Diksha K. Bambulkar, Simarn S. Sahare, improved prediction accuracy
Learning Isolation Forest, One-Class SVM, Local
Mrunali S. Shende, Prof. Vaishnavi Ganesh
Outlier Factor
review includes multiple algorithms such as
Omar Enzo Santangelo, Vito Gentile, LSTM achieves high accuracy in several cases,
Machine Learning and Prediction of Infectious Long Short-Term Memory, Random Forest,
Stefano Pizzo, Domiziana Giordano, and with specific examples like a 93% accuracy for
Diseases: A Systematic Review Support Vector Machine, Autoregressive
Fabrizio Cedrone. COVID-19
Integrated Moving Average

LASVM (Least Angle Support Vector

Fahmi Ben Rejab, Kaouther Nouira, and
Health Monitoring Systems Using Machine Machines)
Abdelwahed Trabelsi
Learning Techniques ISVM (Incremental Support Vector Machines)
K-Prototypes Clustering
Support Vector Machine (SVM),
Long Short-Term Memory (LSTM), CNN achieving 81% accuracy for pneumonia
Hafsa Habehh and Suril Gohel Convolutional Neural Network (CNN), detection on chest X-rays.
Machine Learning in Healthcare
Decision Trees (DT), DBN reaching 91.76% accuracy for
Deep Belief Networks (DBN), Alzheimer's disease detection using MRI.
Recurrent Neural Networks (RNN)

Effective Heart Disease Prediction Using Hybrid Senthilkumar Mohan, Chandrasegar Hybrid Random Forest with Linear Model
88.7%
Machine Learning Techniques Thirumalai, Gautam Srivastava (HRFLM)
5
Problem Statement
• Traditional methods are unable to adequately provide proactive and tailored health
monitoring, which is necessary due to the rising incidence of chronic diseases and
lifestyle-related health problems.

• Predictive analytics is frequently not used by current healthcare solutions to foresee

personal health hazards and suggest prompt remedies

• The inability of many personalized health monitoring systems to integrate machine

learning limits their capacity to extract precise and useful insights from complicated
medical data.

• Effective machine learning implementation in healthcare is hampered by the difficulty of

managing big, heterogeneous, and delicate health datasets.

• Research on how machine learning-powered predictive analytics might be customized to

account for individual differences in health conditions and preferences is lacking.
6
Methodology
Involves
 Several structured steps
 Combining data acquisition
 Preprocessing
 Model development

7
Methodology Diagram

Testing
(30%)

8
Methodology (Data Collection)
Data set collected from
 Kaggle
 Which had 4,920 occurrences
 133 Features.
 Each feature represents a distinctive symptom, recorded as
binary values (1for presence, 0 for absence)
 The dataset associates these symptoms with 41 illnesses.

9
Methodology (visual of training dataset)

10
Figure: 01
Methodology (Data Pre-processing)
 Handling the missing value:
If any dataset contains missing value, it may create problem in machine learning
classification model. Our dataset contains some missing value, we used python Scikit
learn library to handle the missing value
 Encoding Categorical Data:
Encoded into numerical using label encoding.
 Feature Labeling:
For the purpose of disease prediction, label encoding was applied to the Prognosis
column to transform disease names into numerical values.

11
Methodology (Data Pre-processing)

Each bar in the bar graph represents

• A distinct disease
• Shows the distribution of diseases in the
dataset.
Dataset is balanced
• With about similar proportions for each of the
41 medical conditions
• This balance reduces bias towards any
particular class during classification

Figure 02: Unique Disease Graph 12

Methodology (Feature Selection)
Recursive Feature Elimination (RFE):
RFE was applied to iteratively remove the least
important features, and thus help the classifiers
perform better with a reduced set of features. This
step also reduced the computational complexity.
Mutual Information:
Mutual information was used as a filter-based
method to assess the relevance of features by
measuring the mutual dependency between the
features and the target variable.
Figure 03: Correlation graph between features and prognosis
13
Methodology (Data Pre-processing)

Figure 04: Important Feature 1 Figure 05: Important Feature 2

14
Methodology (Algorithms)
 Support Vector Classifier (SVC)
 Random Forest Classifier(RF)
 Gradient Boosting Classifier
 Logistic Regression
 Decision Tree
 Multinomial Naïve Bias
 K- Nearest Neighbors(KNN)
 Ensemble Classifier
 Bagging (Bootstrap Aggregating)
 Boosting

15
Methodology (Proposed web Model)

Input
User Disease
Login Health
Interface Name
page Problem

Sign up
Medicine Diet Exercise
suggestion suggestion Suggestion

16
Conclusion

The use of machine learning models has proven effective in personalized health monitoring and
predictive analytics. This study applied various techniques, including Random Forest, K-
Nearest Neighbors, Multinomial Naive Bayes, Support Vector Classifier, Gradient Boosting,
Decision Tree, and Logistic Regression, to predict diseases based on user-input symptoms.

17
Future Work
• Integration of Advanced Machine Learning Techniques
• Expansion of the Dataset
• Improved Model Interpret-ability
• User Interface and Experience Enhancements
• Collaboration with Healthcare Experts
• Validation and Testing in Real-World Settings

18
Reference
[1] Mohammed Badawy, Nagy Ramadan, and Hesham Ahmed Hefny. Healthcare predictive analytics using machine learning
and deep learning techniques: a survey. Journal of Electrical Systems and Information Technology, 10(1):40, 2023.
[2] Shahadat Uddin, Arif Khan, Md Ekramul Hossain, and Mohammad Ali Moni. Comparing different supervised machine
learning algorithms for disease prediction. BMC Medical Informatics and Decision Making, 19(1):281, 2019.
[3] Seema S. Awathare, Samiksha G. Gajbhiye, Diksha K. Bambulkar, Simarn S. Sahare, Mrunali S. Shende, and Vaishnavi
Ganesh. Heart disease prediction using machine learning. International Journal of Advanced Research in Computer and
Communication Engineering, 13(3):473–477, 2024.
[4] Omar Enzo Santangelo, Vito Gentile, Stefano Pizzo, Domiziana Giordano, and Fabrizio Cedrone. Machine learning and
prediction of infectious diseases: A systematic review. Machine Learning and Knowledge Extraction, 5:175–198, 2023. Open
access under CC BY license.
[5] Fahmi Ben Rejab, Nouira Kaouther, and Abdelwahed Trabelsi. Support vector machines versus multi-layer perceptrons
for reducing false alarms in intensive care units. International Journal of Computer Applications, 49:975–8887, 08 2012.
[6] Hafsa Habehh and Suril Gohel. Machine Learning in Healthcare.National Library of Medicine.2021
[7] Senthilkumar Mohan, Chandrasegar Thirumalai, and Gautam Srivastava. Effective heart disease prediction using hybrid
machine learning techniques. IEEE Access, 7:81542–81554, 2019.

19
THANK
YOU