IoT based Illness Prediction System using Machine

Learning

A Seminar Report
Submitted to the APJ Abdul Kalam Technological University
in partial fulfillment of requirements for the award of degree

Bachelor of Technology
in
Computer Science and Engineering
by
Faiz
EKC21CS017

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

ERANAND KNOWLEDGE CITY TECHNICAL CAMPUS
MANJERI
October 2024
 DEPT. OF COMPUTER SCIENCE & ENGINEERING ERANAD
KNOWLEDGE CITY TECHNICAL CAMPUS MANJERI
2024-25

CERTIFICATE

This is to certify that this Seminar Report titled IoT based Illness Prediction
System using Machine Learning is the bonafide record of the work done by Faiz
(EKC21CS017) of seventh semester, Computer Science and Engineering, Eranad
Knowledge City Technical Campus, towards the partial fulfillment of the requirements
for the award of the Degree of Bachelor of Technology by the A P J Abdul Kalam
University.

Ms. Anu K S Ms. Shyni K

(Seminar Guide) (Seminar Coordinator)
HOD Asst. Professor
Dept.of CSE Dept.of CSE
EKCTC EKCTC
Manjeri Manjeri
 DECLARATION

I Faiz hereby declare that the seminar report IoT based Illness Prediction System
using Machine Learning , submitted for partial fulfillment of the requirements for the
award of degree of Bachelor of Technology of the APJ Abdul Kalam Technological
University, Kerala is a bonafide work done by me under supervision of Ms. Anu K S .
This submission represents my ideas in my own words and where ideas or words
of others have been included, I have adequately and accurately cited and referenced
the original sources.
I also declare that I have adhered to ethics of academic honesty and integrity
and have not misrepresented or fabricated any data or idea or fact or source in my
submission. I understand that any violation of the above will be a cause for disciplinary
action by the institute and/or the University and can also evoke penal action from the
sources which have thus not been properly cited or from whom proper permission has
not been obtained. This report has not been previously formed the basis for the award
of any degree, diploma or similar title of any other University.

EKCTC Manjeri Faiz

07-10-2024
Abstract

The increasing integration of Internet of Things (IoT) and Machine Learning (ML)
in healthcare has revolutionized real-time health monitoring and illness prediction.
This report explores the development of an IoT-based Illness Prediction System
that utilizes wearable devices to continuously track vital health data such as blood
pressure, pulse rate, and glucose levels. By employing K-Nearest Neighbor (KNN)
and Support Vector Machine (SVM) algorithms, the system processes the collected
data to predict potential health risks and provide preventive care. The proposed
framework focuses on data preprocessing, classification, and real-time analysis to
enhance the accuracy of predictions. The system has the potential to reduce healthcare
costs, improve early detection of illnesses, and optimize workplace health monitoring.
Evaluation metrics such as accuracy, precision, and recall were used to assess the
performance of the algorithms, demonstrating their effectiveness in classifying health
data. This report also outlines the challenges, benefits, and future scope of the system,
which includes enhancing sensor accuracy and incorporating advanced AI techniques
for better predictive outcomes.

i
Acknowledgement

I take this opportunity to express my deepest sense of gratitude and sincere thanks
to everyone who helped me to complete this work successfully. I would remember
with grateful appreciation, the encouragement and support rendered by Dr. Sobha P,
Principal, Eranad Knowledge City, Manjeri. I express my sincere thanks to Ms Anu
K S, Head of Department, Computer Science and Engineering, Eranad Knowledge
City Technical Campus Manjeri for providing me with all the necessary facilities and
support.
I would like to express my sincere gratitude to Ms. Shyni. and depart-
ment of Computer Science and Engineering, Eranad Knowledge City Technical
Campus Manjeri for their support and co-operation.
I would like to place on record my sincere gratitude to my seminar guide
Ms. Anu K S , HOD of department of Computer Science and Engineering, Eranad
Knowledge City Technical Campus for the guidance and mentorship throughout the
course.
Finally I thank my family, and friends who contributed to the successful fulfilment
of this seminar work.

Faiz

ii
Contents

Abstract i

Acknowledgement ii

List of Figures v

1 Introduction 1

2 Literature Survey 2
2.1 A Novel Machine Learning Technique for Diabetic Prediction in IoT-
based Healthcare Monitoring System . . . . . . . . . . . . . . . . . 3
2.1.1 Methodology: . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1.2 Advantages: . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1.3 Disadvantages: . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Health Disease Prediction Using Machine Learning and Internet of
Things . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2.1 Methodology: . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2.2 Advantages: . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2.3 Disadvantages: . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.3 Intelligent IoT-based Combined Crop-type and Disease Prediction
System with Machine Learning . . . . . . . . . . . . . . . . . . . . 4
2.3.1 Methodology: . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.3.2 Advantages: . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3.3 Disadvantages: . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4 Improved IoT for Health Behaviour System Based on Machine Learn-
ing Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

iii
 2.4.1 Methodology: . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4.2 Advantages: . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4.3 Disadvantages: . . . . . . . . . . . . . . . . . . . . . . . . . 5

3 Proposed System 6
3.1 Machine Learning Framework . . . . . . . . . . . . . . . . . . . . . 6
3.1.1 Data Segmentation: . . . . . . . . . . . . . . . . . . . . . . . 6
3.1.2 Dimensionality Reduction: . . . . . . . . . . . . . . . . . . . 7
3.1.3 Supervised Learning Approach: . . . . . . . . . . . . . . . . 7
3.2 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.2.1 Loading and Preprocessing: . . . . . . . . . . . . . . . . . . 8
3.3 Data Preprocessing and Feature Engineering . . . . . . . . . . . . . . 8
3.4 K-Nearest Neighbor (KNN) . . . . . . . . . . . . . . . . . . . . . . 9
3.4.1 Key Features of KNN: . . . . . . . . . . . . . . . . . . . . . 9
3.4.2 How KNN Works in Illness Prediction: . . . . . . . . . . . . 10
3.4.3 Advantages of KNN: . . . . . . . . . . . . . . . . . . . . . . 10
3.4.4 Disadvantages of KNN: . . . . . . . . . . . . . . . . . . . . 10
3.5 Support Vector Machine (SVM) . . . . . . . . . . . . . . . . . . . . 11
3.5.1 Key Features of SVM: . . . . . . . . . . . . . . . . . . . . . 11
3.5.2 How KSVM Works in Illness Prediction: . . . . . . . . . . . 11
3.5.3 Advantages of SVM: . . . . . . . . . . . . . . . . . . . . . . 11
3.5.4 Disadvantages of SVM: . . . . . . . . . . . . . . . . . . . . 12

4 Results and Discussion 13

4.1 Evaluation Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

5 Conclusion and Future Scope 15

References 16

iv
List of Figures

3.1 System Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . 7

3.2 Loading and Preprocessing . . . . . . . . . . . . . . . . . . . . . . . 9

4.1 Dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.2 Relevant Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

v
Chapter 1

Introduction

In the era of technological advancements, healthcare systems are evolving rapidly

with the integration of Internet of Things (IoT) and Machine Learning (ML). These
technologies are reshaping how health data is collected, processed, and analyzed.
The development of IoT-based illness prediction systems focuses on the utilization
of wearable technology to monitor real-time health conditions. Through the constant
collection of vital signs such as pulse rate, blood pressure, glucose levels, and more,
these systems aim to provide continuous health monitoring. In this report, we explore
the implementation of an IoT-based Illness Prediction System using ML algorithms
like K-Nearest Neighbor (KNN) and Support Vector Machine (SVM) to classify
health data, predict potential illnesses, and support preventive care.

1
Chapter 2

Literature Survey

1. A Novel Machine Learning Technique for Diabetic Prediction in IoT-based

Healthcare Monitoring System

2. Health Disease Prediction Using Machine Learning and Internet of Things

3. Intelligent IoT-based Combined Crop-type and Disease Prediction System

with Machine Learning

4. Improved IoT for Health Behaviour System Based on Machine Learning

Model

2
2.1 A Novel Machine Learning Technique for Dia-
betic Prediction in IoT-based Healthcare Monitor-
ing System
Authors: Dr. S. Aravinth Vijay Jesuraj, M. Sridharan, R. Ganesh, Dr. Vivek
Ravishankar Dubey, Vinda Manjramkar
This study introduces the development of a system that uses Tuna Swarm
Optimization-Aided Neural Classifier (TSO-NN) for diabetes prediction. The
system utilizes IoT-based wearable devices to collect patient data such as glucose levels
and processes the data through ML techniques. The model’s performance is evaluated
based on accuracy, precision, and sensitivity.

• 2.1.1 Methodology:

Data is gathered from wearable sensors and classified using a neural network
enhanced by optimization techniques. Metrics like F1-score are used for
evaluation.

• 2.1.2 Advantages:

Improved accuracy and real-time monitoring capabilities.

• 2.1.3 Disadvantages:

Dependency on device accuracy and complexity in system integration.

2.2 Health Disease Prediction Using Machine Learning

and Internet of Things
Authors: Patil Rahul Ashokrao, Derle Deepak Radhakrishna, Mukesh Sharma,
Sandeep Bhad

3
 This paper discusses the use of IoT for disease prediction. Decision Trees, KNN,
and Naı̈ve Bayes algorithms are applied to health data collected from IoT devices for
early detection of diseases. The system is designed for real-time data processing using
cloud and fog computing.

• 2.2.1 Methodology:

IoT devices collect health data, which is processed in the cloud for real-time
predictions using ML classifiers.

• 2.2.2 Advantages:

Cost-effective diagnostics and reduced workload for healthcare providers.

• 2.2.3 Disadvantages:

Data privacy concerns and reliance on the quality of collected data.

2.3 Intelligent IoT-based Combined Crop-type and Dis-

ease Prediction System with Machine Learning
Authors: Md. Ishan Arefin Hossain, Syeda Mahin Binta Haque, Ahmed Kiser, Israt
Jahan Mitu
This study extends IoT’s use into agriculture, combining Raspberry Pi, Azure
IoT, and various sensors to predict crop diseases. Though the domain differs, the
integration of ML techniques like Deep Learning shows the versatility of IoT in health
and environmental applications.

• 2.3.1 Methodology:

Data collected from agricultural sensors is processed using machine learning to

predict crop diseases.

4
 • 2.3.2 Advantages:

High accuracy (94.18%) and cost-effective setup using low-cost sensors.

• 2.3.3 Disadvantages:

Complex setup and resource limitations due to hardware dependencies.

2.4 Improved IoT for Health Behaviour System Based

on Machine Learning Model
Authors: Dr. Anurag Shrivastava, Dr. Midhun Chakkaravarthy, Dr. Mohd Asif Shah
This research proposes a multi-layered ML model for health behavior prediction,
comparing algorithms like KNN and SVM. The model aims to improve patient
monitoring and diagnostic accuracy through real-time data analysis.

• 2.4.1 Methodology:

Multiple layers of ML algorithms are used to classify health behavior data with
a focus on accuracy and precision.

• 2.4.2 Advantages:

Enhanced diagnostic capabilities and real-time monitoring.

• 2.4.3 Disadvantages:

High implementation costs and potential privacy issues.

5
Chapter 3

Proposed System

The proposed system focuses on using IoT-based wearable devices to monitor

employees’ health in real time, especially in workplace settings. The system is
designed to predict illness using machine learning techniques and provide preventive
care. This section elaborates on the methodologies utilized in the system.

3.1 Machine Learning Framework

A Machine Learning (ML) Framework is the foundation for developing intelligent
systems that can process large datasets to uncover patterns, classify data, and make
predictions. In the case of the IoT-based Illness Prediction System, the ML
framework is essential for handling vast amounts of health data gathered from wearable
devices.

3.1.1 Data Segmentation:

The data collected from wearables are divided into training and testing datasets.
Training data is used to develop the model by teaching the algorithm to recognize
patterns, while testing data is used to validate the model’s accuracy. Splitting data
ensures that the model generalizes well and performs accurately on unseen data.

6
 Figure 3.1: System Architecture

3.1.2 Dimensionality Reduction:

Given that health data can include numerous parameters (such as blood pressure,
glucose levels, pulse rate, etc.), dimensionality reduction techniques are applied to
optimize performance. This process involves reducing the complexity of the dataset
by eliminating irrelevant or redundant features, which in turn improves the efficiency
and accuracy of ML algorithms.

3.1.3 Supervised Learning Approach:

The system employs supervised learning, where the input data (e.g., health parameters)
is labeled with output categories (e.g., healthy or unhealthy conditions). Supervised
learning is advantageous because the algorithms learn from labeled data and make
more precise predictions.

7
3.2 Data Collection
Wearable Devices play a critical role in the IoT-based health monitoring system.
These devices collect a wide range of physiological data from users, including:

• Blood Pressure: Monitored continuously to detect hypertension or hypotension,

which can indicate potential health risks.

• Pulse Rate: Tracks heart rate over time to identify abnormal heart rhythms or
early signs of cardiovascular issues.

• Blood Glucose Levels: Particularly crucial for diabetes monitoring, glucose

levels provide insight into metabolic health.

• Temperature: Continuous tracking of body temperature can help in early

detection of fever or infection.

The data is transmitted wirelessly from the wearable devices to a central server or
cloud platform, where it is stored, processed, and analyzed using the machine learning
framework.

3.2.1 Loading and Preprocessing:

Before applying ML techniques, the collected data undergoes preprocessing. Prepro-

cessing includes cleaning the data, handling missing values, and normalizing data to
ensure that all values fall within a comparable range. This step ensures the data is in
optimal form for analysis and minimizes errors during classification.

3.3 Data Preprocessing and Feature Engineering

Data preprocessing is a crucial step in ensuring that the data is in a form suitable for
analysis. This includes:

• Data Cleaning: Removing or filling missing values to avoid errors in the

algorithms.

8
 Figure 3.2: Loading and Preprocessing

• Normalization: Standardizing the range of independent variables (features) so

that they contribute equally to the results.

• Feature Engineering: Creating new features or modifying existing ones to

better capture the underlying patterns in the data. For example, combining pulse
rate and temperature to derive a new health metric.

3.4 K-Nearest Neighbor (KNN)

K-Nearest Neighbor (KNN) is a classification algorithm that is both simple and
effective, especially in the healthcare domain. It is widely used for classification
problems where the objective is to assign a new data point to one of the predefined
categories.

3.4.1 Key Features of KNN:

• Instance-based Learning: KNN belongs to a group of instance-based learn-

ers, meaning it does not create an explicit model. Instead, it compares new data
points directly to the training examples.

• Distance Calculation: KNN uses Euclidean distance to measure how similar

or different two data points are. For example, if the system is analyzing glucose
levels and heart rate data, KNN will calculate the distance between the current

9
 data point (new reading) and previous data points (labeled readings). The
algorithm looks for the closest points to the new data.

• K Parameter: The value of K (number of nearest neighbors to consider) is

critical. A low value of K (e.g., K=3) might lead to noisy predictions, while a
high value of K may overgeneralize. Therefore, the choice of K is based on trial
and error to achieve the best accuracy.

3.4.2 How KNN Works in Illness Prediction:

1. The system receives new health data (e.g., a person’s current blood pressure and
pulse rate).

2. It compares this new data to the historical dataset (labeled with healthy or
unhealthy conditions).

3. Based on the closest matching data points, the algorithm predicts whether the
person is at risk of illness (e.g., hypertension or heart disease).

3.4.3 Advantages of KNN:

• Simplicity: Easy to implement and understand.

• Effectiveness: Performs well in real-time scenarios where quick classifications

are needed.

3.4.4 Disadvantages of KNN:

• Computationally Intensive: As the size of the dataset grows, the time required
to find the nearest neighbors increases.

• Sensitive to Irrelevant Features: KNN may become less accurate if irrelevant

features are present in the data.

10
3.5 Support Vector Machine (SVM)
Support Vector Machine (SVM) is a supervised learning algorithm that is com-
monly used for classification tasks in the healthcare domain due to its high perfor-
mance in handling complex and high-dimensional data.

3.5.1 Key Features of SVM:

• Hyperplane: SVM creates a decision boundary called a hyperplane, which

separates different classes of data. In the health monitoring system, this
hyperplane might separate ‘healthy’ from ‘unhealthy’ individuals based on
parameters such as blood glucose levels or pulse rates.

• Maximizing Margin: The algorithm aims to find the widest possible margin
between the two classes. This margin is the distance between the hyperplane and
the closest data points from each class, known as support vectors.

• Non-linear Classification: If the data is not linearly separable (i.e., it cannot be

split by a straight line), SVM uses a technique called the kernel trick to project
the data into higher-dimensional space where it becomes easier to separate.

3.5.2 How KSVM Works in Illness Prediction:

1. The system receives health data from wearable devices (e.g., temperature,
glucose levels).

2. The SVM algorithm constructs a hyperplane that best separates healthy individ-
uals from those at risk of illness.

3. If the new data point falls on one side of the hyperplane, the system classifies it
as ‘healthy’; if it falls on the other side, it is classified as ‘unhealthy’ or at risk.

3.5.3 Advantages of SVM:

• High Accuracy: SVM is known for its ability to perform well in high-
dimensional spaces and deliver precise predictions.

11
 • Effective with Small Datasets: It works well even when the dataset is not very
large, making it suitable for healthcare settings where data collection might be
limited.

3.5.4 Disadvantages of SVM:

• Complexity: The implementation of SVM can be more complex compared to

simpler algorithms like KNN, particularly when selecting the right kernel.

• Time-Consuming: Training SVM can be time-consuming, especially for large

datasets.

12
Chapter 4

Results and Discussion

4.1 Evaluation Metrics

The effectiveness of both KNN and SVM models is evaluated using the following
metrics:

• Accuracy: The percentage of correct predictions made by the model.

• Precision: Measures the accuracy of the positive predictions made by the model
(e.g., how many individuals predicted to have an illness actually have it).

• Recall: Measures the model’s ability to identify all positive cases (e.g., how
many individuals who have an illness were correctly identified).

• F1-Score: A harmonic mean of precision and recall, providing a balanced

measure when there is an uneven class distribution.

By using these metrics, the system ensures that the predictions made by the KNN and
SVM algorithms are reliable and actionable for healthcare providers and individuals.
The performance of the proposed system was evaluated based on the accuracy of
the KNN and SVM algorithms. The results showed that both algorithms performed
well in classifying health data, with SVM showing slightly higher accuracy due to
its ability to handle high-dimensional data more efficiently. The system successfully
demonstrated the potential for early illness detection and preventive care through
continuous monitoring.

13
 Figure 4.1: Dataset

Figure 4.2: Relevant Results

• Accuracy of KNN: 8̃9%

• Accuracy of SVM: 9̃2%

• Key Findings: The integration of wearable IoT devices with ML models

enhances real-time health monitoring and illness prediction.

14
Chapter 5

Conclusion and Future Scope

In conclusion, the IoT-based illness prediction system using machine learning offers
a promising solution for real-time health monitoring in workplace environments. The
integration of wearable devices with ML algorithms like KNN and SVM allows for
early illness detection and preventive care, reducing medical leave and healthcare costs.
Future Scope:

1. Improved sensor accuracy for more reliable health monitoring.

2. Integration of advanced AI techniques like Deep Learning to enhance prediction

models.

3. Expansion of the system for large-scale health monitoring across various

industries.

4. Use of predictive analytics for preventive healthcare.

15
References

1. S. Aravinth Vijay Jesuraj, M. Sridharan, et al., ”A Novel Machine Learning

Technique for Diabetic Prediction in IoT-based Healthcare Monitoring System,”
2023.

2. Patil Rahul Ashokrao, Derle Deepak Radhakrishna, ”Health Disease Prediction

Using Machine Learning and Internet of Things,” 2023.

3. Md. Ishan Arefin Hossain, Syeda Mahin Binta Haque, ”Intelligent IoT-based
Combined Crop-type and Disease Prediction System with Different Machine
Learning Techniques,” 2023.

4. Dr. Anurag Shrivastava, Dr. Midhun Chakkaravarthy, et al., ”Improved IoT for
Health Behaviour System Based on Machine Learning Model,” 2022.

5. J. Hayano, H. Yamamoto, I. Nonaka et al., “Quantitative detection of sleep apnea

with wearable watch device,” PLoS ONE, vol. 15, Article ID e0237279, 2020.

6. F. Delmastro, F. D. Martino, and C. Dolciotti, “Cognitive training and stress

detection in MCI frail older people through wearable sensors and machine
learning,” IEEE Access, vol. 8, Article ID 65573, 2020.

7. Khan Y, Ostfeld AE, Lochner CM, Pierre A, Arias AC. “Monitoring of vital
signs with flexible and wearable medical devices,” Advanced Materials, vol. 28,
no. 22, pp. 4373-4395, 2016.

8. Guze PA, “Using technology to meet the challenges of medical education,”

Transactions of the American Clinical and Climatological Association, vol. 126,
pp. 260-270, 2015.

16
 9. S. Seneviratne, Y. Hu, T. Nguyen et al., “A survey of wearable devices and
challenges,” IEEE Communications Surveys & Tutorials, vol. 19, no. 4, pp.
2573-2620, 2017.

10. M. Chan, D. Estève, J.-Y. Fourniols, C. Escriba, and E. Campo, “Smart

wearable systems: current status and future challenges,” Artificial Intelligence
in Medicine, vol. 56, no. 3, pp. 137–156, 2012.

11. S. S. Kale and D. Bhagwat, “Highly secured IoT-based healthcare system for
elderly people using body sensor network,” International Journal of Innovative
Research in Science Engineering and Technology, vol. 5, no. 10, 2016.

12. N. Hussain, “IEEE 802.15.6 standard in wireless body area networks from a
healthcare point of view,” IEEE Sensors Journal, vol. 16, no. 5, pp. 1368-1376,
2016.

17