You are on page 1of 12

Similarity Report ID: oid:27005:52404602

PAPER NAME AUTHOR

BIBA Enhancing Heart Disease Predictio 1 2


n with a Hybrid Model Combining Decisio
n Tree, Logistic Regres

WORD COUNT CHARACTER COUNT

3059 Words 18320 Characters

PAGE COUNT FILE SIZE

7 Pages 299.2KB

SUBMISSION DATE REPORT DATE

Feb 17, 2024 11:31 AM GMT+5:30 Feb 17, 2024 11:32 AM GMT+5:30

20% Overall Similarity


The combined total of all matches, including overlapping sources, for each database.
15% Internet database 13% Publications database
Crossref database Crossref Posted Content database
17% Submitted Works database

Excluded from Similarity Report


Bibliographic material

Summary
Enhancing Heart Disease Prediction with a Hybrid Model
2
Combining Logistic Regression, Decision Tree and Artificial
Neural Network
Nura Muhammad Sani1
MCA Sem – II
4
Datta Meghe Institute of Higher Education & Research, Sawangi (M)
Email – nuramuhammad07061@gmail.com

Ms. Reena S. Satpute2


Assistant Professor
School of Allied Sciences, Faculty of Science & Technology
4
Datta Meghe Institute of Higher Education & Research, Sawangi (M)
Email – reenasatpute2017@gmail.com

Abstract
Heart disease, a leading cause of global mortality, approaches could enable accurate, cost-effective early
necessitates accurate prediction for timely prediction.
intervention. This
2
study proposes a hybrid model
amalgamating Logistic Regression, Decision Tree and 2. Related Work
Artificial Neural Network (ANN) algorithms to Heart disease weakens the heart’s ability to pump
enhance heart disease prediction. Using a Kaggle blood, leading to health issues like shortness of breath
dataset comprising 1025 patient records with 14 and fatigue [5]. It is prevalent globally, especially in
features, including age, sex, chest pain, and cholesterol developing countries, where limited diagnostic tools
levels, the hybrid model achieved an impressive 88% pose challenges for early diagnosis and treatment [5].
precision. This outperforms individual models, with Accurate early prediction enables effective
DT achieving 99% accuracy, LR with 80%, and ANN intervention, improving outcomes.
with 86%. Evaluation metrics demonstrate Machine learning techniques like logistic regression
competitive performance, affirming the hybrid model [6], random forests [7] and deep belief networks [9]
as a robust tool for cardiovascular ailment prediction. have been explored for heart infection prediction using
The study underscores the efficacy of combining datasets from UCI repository and hospitals [5][10].
diverse algorithms, leveraging their strengths for more Classification accuracy ranges widely, from 77% [6]
effective predictive modeling in cardiovascular health to 100% [3], reflecting differences in algorithms and
assessment. data used. More complex ensemble approaches tend to
outperform individual models. For example,
Keywords : Decision Tree, Logistic Regression, Alizadehsani [11] combined multiple classifiers using
Artificial Neural Network, hybrid model an ensemble method, achieving 90% accuracy. Ruzzo-
Tompa’s [9] optimally configured deep belief network
1.Introduction attained 94.61% accuracy. Feature engineering and
selection techniques also boost performance [10].
The heart circulates blood, supplying nutrients and However, issues like overfitting [13] and suboptimal
oxygen to bodily organs [1]. Heart disease impairs this model tuning [17] persist. Most studies test individual
function, affecting the brain, lungs, liver, and kidneys algorithms rather than hybrid approaches that could
[1]. 33
It causes cognitive and respiratory issues [2]. As mitigate limitations. As Jabbar [13] demonstrated,
per the World Health Organization, heart disease minimizing training error sometimes increases testing
accounts for 31% of annual deaths globally [3]. errors. Ahmed [17] achieved only 80.32% validation
Delayed diagnosis leads to poor outcomes and accuracy despite testing multiple models.
mortality [3]. Machine learning models show promise Opportunities exist to advance accuracy further
for early prediction to improve prognosis [4]. through larger, more diverse datasets [5], ensembles
However, traditional models struggle with complexity and hybrid methods [11], improved feature selection
while deep learning models require extensive data and [10], and better hyper parameter optimization and
risk overfitting [4]. A hybrid model combining both cross-validation [10]. Combining complementary
techniques in a tuned, thoroughly validated hybrid positives. These metrics provide a comprehensive
approach could yield both high accuracy and understanding of the models' strengths and
reliability in diagnosis. weaknesses.

3. Methodology 3.5.1 Evaluation Metrics


22
The learning aims to predict heart disease possibility Models are evaluated using four standard metrics:
using computerized prediction, benefiting medical precision, precision, recall, and F1 score. In these
3
professionals and patients. To accomplish this measurements, TP represents true positives, TN
objective, a systematic approach was adopted. This represents true negatives, FP represents false
process is illustrated in Figure 1 below: positives, and FN represents false negatives.
Equations 1, 2, 3, and 4 present the mathematical
34
3.1 Data Collection expressions of classification accuracy (CA), precision,
This study uses a Kaggle dataset of 1025 patient recall, and F1 score, respectively. These calculations
records with 14 distinct features related to heart are performed on experimental data.
disease risk factors. The distribution of the data set is
shown in the figure below:
START
3.2 Data Preprocessing
Before starting to train the model, the dataset went
20
over and done with preprocessing to ensure quality Data
and consistency. This involves correcting missing collectio
values, normalizing features, and dividing the data into
39
training and testing sets. Missing values were Preprocessing
addressed through standard scalar techniques,
preserving the integrity of the dataset. Feature
Model selection
normalization was conducted to bring all features to a
common scale, enabling fair comparison among
different models.
18 Decision Artificial Logistic
3.3 Data Splitting Tree Neural regression
The dataset was distributed into training and testing Network
sets, with a ratio of 80:20. Stratified sample was
employed to ensure balanced class distribution. Models training &
testing
3.4 31
Model Training
The heart disease prediction work was tackled using Hybrid Model
2
three distinct machine learning models: Logistic
Regression, Decision Tree 24
and Artificial Neural
Network (ANN). Each model was trained on the Model
training data using their respective algorithms. The Evaluation
Decision Tree model leverages a hierarchical structure
to partition feature space; Logistic Regression Result Analysis
establishes a lined relationship between features and
the target, while the ANN16model comprises multiple
interconnected layers to capture complex patterns STOP
within the data. Hybrid model is proposed, which
combines the predictions from all three individual 14
models to leverage their complementary strengths. Figure 1: Block Diagram of Proposed Model
25 (𝑇𝑃+𝑇𝑁)
3.5 Performance Evaluation and Evaluation 𝐴𝑐𝑐𝑢𝑟𝑎𝑐𝑦 = . . 1
(𝑇𝑃+𝐹𝑃+𝑇𝑁+𝐹𝑁)
Metrics 𝑇𝑃
To evaluate the models' predictive performance, 𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛 = . . 2
𝑇𝑃+𝐹𝑃
numerous metrics were employed comprising 𝑇𝑃
7 𝑅𝑒𝑐𝑎𝑙𝑙 = . . 3
accuracy, precision, recall, and F1-score. Accuracy 𝑇𝑃+𝐹𝑁
𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛∗𝑅𝑒𝑐𝑎𝑙𝑙
quantifies13 correctness of the predictions, precision 𝐹1 − 𝑆𝑐𝑜𝑟𝑒 = 2( ). 4
𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛+𝑅𝑒𝑐𝑎𝑙𝑙
measures proportion of true positive predictions out of
all positive predictions. F1-score provides a harmonic 4. Result and discussion
7
mean of precision and recall. Recall evaluates the The research employed Google Colab on a system
proportion of true positive predictions out of all actual featuring an Intel(R) Pentium(R) processor with 8 GB
RAM. The initial dataset comprised 1,025 rows and 14
attributes. However, following data cleaning and
preprocessing, it was reduced to 820 rows and 13
attributes. The study involves various stages, starting
with data collection, followed by preprocessing. The
subsequent steps include model selection, which
12
encompasses decision tree, artificial neural network,
and logistic regression. The models undergo training
and testing, culminating in a hybrid model. Evaluation
involves19assessing the results and analyzing them,
utilizing metrics like precision, recall, F1 score, and Figure 3: Correlation Matrix
the area under the ROC curve. For algorithm
implementation, logistic regression, decision trees, 4.2 Data Preprocessing
artificial neural networks, and hybrid models are The data was checked for missing values using
8 data.isnull().sum() and no missing values were found.
employed. Performance metrics such as precision,
recall, F1 score, and area under the ROC curve are data.duplicated().sum() showed there were no
28 duplicate rows. StandardScaler from
used to evaluate the models. The dataset is partitioned
into 80% for training and 20% for testing. sklearn.preprocessing was used to standardize the
30
heart disease dataset features to have mean 0 and
4.1 Dataset standard deviation 1. Standardization helps improve
4.1.1 Data Description model performance and stability.
The dataset is made up of over one thousand and
twenty five (rows) and fourteen features (columns). Its
15
features consisting of age, sex, Chest pain(cp),
trestbps, chol, fbs, restecg, thalach, exang, oldpeak,
slope, ca, thal, and target which indicates/classify
heart disease exist or not. Dataset was collected from
the largest community of data science in the globe,
which is www.Kaggle.com

4.1.2 Data Balance


36
This is used to understand the balance of the data. As Figure 4: Count of the number of missing values in
shown in Figure 2 below, the data between cardiac (1) each column of a Data Frame
and non-cardiac (0) diseases are not balanced.
4.3 Training Process

The model training commenced with 21


the fit method,
ensuring specification of training (X_train, y_train)
and validation data (X_test, y_test). Four models—
Decision Tree, Logistic Regression, Artificial Neural
Network (ANN), and Hybrid Model were employed
38
for heart disease prediction. Decision Tree maximizes
class separation through recursive feature-based splits.
Logistic Regression estimates feature coefficients,
Figure2: Data Balance modeling heart disease probability with a logistic
29
function. ANN learns iteratively, adjusting weights
4.1.3 Correlation Matrix and biases via back propagation. The Hybrid Model
combines individual predictions, capturing diverse
The correlation matrix was used to determine the patterns and relationships for a comprehensive heart
relationship or strong correlations with the disease prediction approach.
classification row. As shown in figure 3 bellow:
4.4 Result of Artificial Neural Network
In table 1 below, The study trained an artificial neural
network for heart disease detection over 50 epochs.
Results show training time, step loss, and validation
loss trends indicating improving model accuracy and
decreasing loss. Both training and validation
accuracies were relatively high, suggesting good
learning and generalization. Validation metrics help
determine potential overfitting if validation loss
increases while training loss decreases. Accuracy 0.86 205
Macro Avg 0.86 0.86 0.86 205
6
Table 1: Result of Artificial Neural Network Wt. Avg 0.86 0.86 0.86 205
No Training Time Step 1
loss Accuracy 4.5.3 Logistic Regression Classification report
In Table 4, the logistic regression model shows
moderate performance for heart disease prediction,
1. 1s 2ms/step 0.5692 0.7293 with varying precision and recall per class. It achieves
2. 0s 2ms/step 0.4863 0.7939 80% accuracy indicating correct predictions
5
for 80%
37
3. 0s 2ms/step 0.4278 0.8293 of cases. The F1-scores demonstrate balanced trade-
4. 170s 2ms/step 0.3843 0.8512 offs between precision and recall. Macro and weighted
5. 0s 2ms/step 0.3538 0.8622 average metrics account for class imbalance and also
. . show moderate performance. Overall the model
48.0s 2ms/step 0.1745 0.9317 achieves even predictiveness across outcomes.
49.0s 2ms/step 0.1718 0.9329 10
50.0s 2ms/step 0.1693 0.9341 Table 4: Logistic Regression Classification Report:
Precision Recall F1-Score Support
0 0.85 0.72 0.78 102
4.5 Classification report for each model 1 0.76 0.87 0.81 103
A classification 23report evaluates a model's
performance using precision, recall, F1-score, and Accuracy 0.8 205
support for each class. It consists of various Macro
6
Avg 0.8 0.79 0.79 205
components for each model. Let's break down the Wt. Avg 0.8 0.8 0.79 205
components of the classification report for each
model: 4.5.4 Hybrid Model Classification report
In Table 5, The hybrid model demonstrates good
4.5.1 Decision tree Classification report balanced performance for heart disease prediction,
The decision tree model shows excellent performance with 90% and 86% precision, 85% and 90% recall for
for heart disease classification, with 97-100% each class, and F1-scores of 0.87. It achieves an
precision, ~99% recall for both classes, and F1-scores overall accuracy of 88% on a balanced test set. Macro
of 0.99, demonstrating well-balanced predictive and weighted average metrics show the model has
capabilities. It achieves 99% overall accuracy on balanced performance despite class imbalance.
balanced class data. High macro and weighted average 9
metrics indicate robust and evenly distributed Table 5: Hybrid Model Classification Report
predictive performance across outcomes. Precision Recall F1-Score Support
1
0 0.9 0.85 0.87 102
Table 2: Decision Tree Classification Report 1 0.86 0.9 0.88 103
Precision Recall F1-Score Support
0 0.97 1 0.99 102 Accuracy 0.88 205
1 1 0.97 0.99 103 Macro Avg 0.88 0.88 0.88 205
5
Wt. Avg 0.88 0.88 0.88 205
Accuracy
11
0.99 205
Macro Avg 0.99 0.99 0.99 205 4.6 Confusion Matrix
Wt. Avg 0.99 0.99 0.99 205 The confusion matrix was used in this study to provide
detailed information about the model's predictions.
4.5.2 Artificial Neural Network Classification
report 4.6.1 Logistic Regression Confusion Matrix
The ANN model in Table 3 achieves a balanced Figure 9 shows that the model correctly predicted
32
classification for heart disease, with an overall 73(71%) true negatives, 90(o.87) true positives,
5
accuracy of 0.86 on a test set, demonstrating a 13(12%) false positives, and 29 (28%) false negatives.
harmonious trade-off between precision and recall, as
reflected in F1-scores and macro/weighted average
16
metrics that account for class imbalance.
1
Table 3: ANN Classification Report
Precision Recall F1-Score Support
0 0.89 0.83 0.86 102
Figure 9: Logistic Regression Confusion Matrix
1 0.84 0.89 0.87 103
4.6.2 Decision Tree Confusion Matrix
Figure 10 shows that the model correctly predicted
102 (100%) true negatives, 100(98%) true positives, 3
(0.029%) false positives, and 0 (0%) false negatives.

Figure13: ROC Curves


Figure 10: Decision Tree Confusion Matrix
4.8 Comparative Result Analysis
4.6.3 Artificial neural network Confusion Matrix
The proposed hybrid model achieves a 0.88 accuracy,
Figure 11 shows that the model correctly predicted 85
outperforming individual logistic regression (0.80)
(83%) true negatives, 92(89%) true positives, 11
and ANN (0.86) models and approaching the 0.99
(10%) false positives, and 017(16%) false negatives.
decision tree benchmark. The logistic regression
model emphasizes recall for the positive class. The
27
ANN model demonstrates balanced 3
precision (0.84),
recall (0.89) and F1-score (0.87) as shown in Table 2

Figure 11: Artificial neural network Confusion Matrix


Table 2: Comparative Result Analysis
Model accura Precis Recall F1Score AUC
4.6.4 Hybrid model Confusion Matrix cy ion
Figure 12 shows that the model correctly predicted 87 Proposed HM
DT
0.88
0.99
0.86
1.0
0.90
0.97
0.88
0.99
0.99
0.99
(85%) true negatives, 92(90%) true positives, 10 (9%) LR 0.8 0.76 0.87 0.81 0.88
ANN 0.86 0.84 0.89 0.87 0.96
false positives, and 017(14%) false negatives. Existing [16] MLP 0.87 0. 88 0. 84 0. 86 0.95
RF 0. 87 0.89 0. 83 0. 86 0.95
DT 0. 86 0.89 0.81 0. 85 0.94
XGB 0.86 0.88 0. 83. 0.86
0.95
[17] SVM 0.92 0.78
0. 86 0.77
NB 0. 85 0.78
LR 0.98 0.76
LightG 0.99 0.72
BM 100 0.92
XGBoo
st
RF
Figure 12: Hybrid model Confusion Matrix [18] Naïve 0.86 0.82 0.87 0.89
Bayes 0.94 0.86 0.94 0.92
SVM& 0.89 0.66 0.81 0.82
4.7 ROC curve XGB 0.95 0.97 0.94 0.95
SVM
The ROC curves depict the models' ability to and DO
XGBoo
distinguish heart disease cases across varying st
thresholds. All models demonstrate excellent
positive/negative class discrimination. The ANN [16], [17] and [18], showing that the model has
model has an AUC of 0.96, performing between the
26 achieved a high percentage of correct predictions.
Decision Tree and Logistic Regression. The Hybrid
model achieves the highest AUC of 0.99, similar to the 5. Conclusion
Decision Tree, indicating it distinguishes classes as 12
The study combined decision tree, logistic regression,
effectively as the best standalone model. and neural network models into a hybrid model to
35
improve heart disease prediction. The hybrid model
achieved higher accuracy (0.88) than individual
models, nearing the 0.99 decision tree accuracy. It
demonstrated good precision (0.86), recall (0.90),
ROC-AUC (0.99), and F1-score (0.88), balancing
identification of positive cases and overall
performance. Overall, the hybrid model exceeded
individual models, providing accurate and balanced
heart disease predictions on par with the decision tree
benchmark.
Prediction Based on Ruzzo–Tompa and Stacked
6. Recommendations: Genetic Algorithm," IEEE Access, vol. 8, pp. 65947-
While the hybrid model shows promising results, 65958, 2020.
further research could explore:
[10] S. Ambesange, A. Vijayalaxmi, S. Sridevi,
 Fine-tuning model parameters to potentially
Venkateswaran and B. S. Yashoda, "Multiple Heart
improve performance.
 Incorporating additional features and Diseases Prediction using Logistic Regression with
advanced neural network architectures. Ensemble and Hyper Parameter tuning Techniques,"
 Exploring ensemble techniques to combine 2020 Fourth World Conference on Smart Trends in
models dynamically based on instance Systems, Security and Sustainability (WorldS4),
characteristics. London, UK, 2020, pp. 827-832, doi:
References 10.1109/WorldS450073.2020.9210404.
[1] “Structure and Function of the Heart," News-
[11] R. Atallah and A. Al-Mousa, "Heart disease
Medical.net, Sep. 5, 2022. [Online]. Available:
detection using machine learning majority voting
https://www.news-medical.net/health/Structure-and-
ensemble method," in 2019 2nd International
Function-of-the-Heart.aspx.
Conference on New Trends in Computing Sciences
[2] C. M. Bhatt, P. Patel, T. Ghetia, and P. L. Mazzeo,
(ICTCS), October 2019, pp. 1-6.
"Effective heart disease prediction using machine
[12] S. Ismaeel, A. Miri, and D. Chourishi, "Using the
learning techniques," Algorithms, vol. 16, no. 2, p. 88,
Extreme Learning Machine (ELM) technique for heart
2023.
disease diagnosis," in Proceedings of the 2015 IEEE
[3]B. Paul and B. Karn, "Heart disease prediction
Canada International Humanitarian Technology
using scaled conjugate gradient backpropagation of
Conference (IHTC2015), Ottawa, ON, Canada, 31
artificial neural network," Soft Computing, vol. 27, no.
May–4 June 2015, pp. 1–3.
10, pp. 6687-6702, 2023
[13] K.H. Miao, J.H. Miao, and G.J. Miao,
[4] B. Sun, W. Cui, G. Liu, B. Zhou, and W. Zhao, "A
"Diagnosing Coronary Heart Disease using Ensemble
hybrid strategy of AutoML and SHAP for automated
Machine Learning," Int. J. Adv. Comput. Sci. Appl.,
and explainable concrete strength prediction," Case
vol. 7, pp. 30–39, 2016.
Studies in Construction Materials, vol. 19, p. e02405,
[14] A. Khan, M. Qureshi, M. Daniyal, and K.
2023.
Tawiah, "A Novel Study on Machine Learning
[5] S. Begum, "A study for predicting heart disease
Algorithm-Based Cardiovascular Disease Prediction,"
using machine learning," Turkish Journal of Computer
Health & Social Care in the Community, 2023.
and Mathematics Education (TURCOMAT), vol. 12,
[15] S. Nashif, M. R. Raihan, M. R. Islam, and H.
no. 10, pp. 4584-4592, 2021.
Imam, "Heart Disease Detection by Using Machine
[6] M. Gudadhe, K. Wankhade, and S. Dongre,
Learning Algorithms and a Real-Time Cardiovascular
”Decision support system for heart disease based on
Health Monitoring System," World Journal of
support vector machine and articial neural network,”
Engineering and Technology, Scientific Research
in Proceedings of International Conference on
Publishing, Jan. 1, 2018. [Online]. Available:
Computer and Communication Technology (ICCCT),
https://doi.org/10.4236/wjet.2018.64057.
pp. 741-745, Allahabad, India, September 2010.
[16] C. M. Bhatt, P. Patel, T. Ghetia, and P. L. Mazzeo,
[7] M. Kavitha, G. Gnaneswar, R. Dinesh, Y. Rohith
"Effective Heart Disease Prediction Using Machine
Sai, and R. Sai Suraj, "Heart disease prediction using
Learning Techniques," Algorithms, vol. 16, no. 2, p.
hybrid machine learning model," in 2021 6th
88, 2023. [Online]. Available:
International Conference on Inventive Computation
https://doi.org/10.3390/a16020088.
Technologies (ICICT), pp. 1329-1333, IEEE, 2021.
[17] K. Karthick, S. Aruna, R. Samikannu, R.
[8] J. P. Li, A. U. Haq, S. U. Din, J. Khan, A. Khan,
Kuppusamy, Y. Teekaraman, and A. R. Thelkar,
and A. Saboor, "Heart Disease Identification Method
"Implementation of a Heart Disease Risk Prediction
Using Machine Learning Classification in E-
Model Using Machine Learning," Computational and
Healthcare," IEEE Access, vol. 8, pp. 107562-107582,
Mathematical Methods in Medicine, Hindawi
2020.
Publishing Corporation, May 2, 2022. [Online].
[9] S. A. Ali, B. Raza, A. K. Malik, A. R. Shahid, M.
Available: https://doi.org/10.1155/2022/6517716.
Faheem, H. Alquhayz, and Y. J. Kumar, "An
[18] U. Nagavelli, D. Samanta, and P. Chakraborty,
Optimally Configured and Improved Deep Belief
"Machine Learning Technology-Based Heart Disease
Network (OCI-DBN) Approach for Heart Disease
Detection Models," Journal of Healthcare February 27, 2022. [Online]. Available:
Engineering, Hindawi Publishing Corporation, https://doi.org/10.1155/2022/7351061.
Similarity Report ID: oid:27005:52404602

20% Overall Similarity


Top sources found in the following databases:
15% Internet database 13% Publications database
Crossref database Crossref Posted Content database
17% Submitted Works database

TOP SOURCES
The sources with the highest number of matches within the submission. Overlapping sources will not be
displayed.

mdpi.com
1 2%
Internet

ebin.pub
2 <1%
Internet

icada2020.nida.ac.th
3 <1%
Internet

ijisae.org
4 <1%
Internet

Glasgow Caledonian University on 2023-08-13


5 <1%
Submitted works

coursehero.com
6 <1%
Internet

Florida International University on 2023-07-24


7 <1%
Submitted works

Manchester Metropolitan University on 2023-05-26


8 <1%
Submitted works

Sources overview
Similarity Report ID: oid:27005:52404602

docs.bioexcel.eu
9 <1%
Internet

ir.juit.ac.in:8080
10 <1%
Internet

Debabrata Swain, Hardeep Patel, Kevin Patel, Vivek Sakariya, Nishtha ...
11 <1%
Crossref

bmcbioinformatics.biomedcentral.com
12 <1%
Internet

medium.com
13 <1%
Internet

ijariit.com
14 <1%
Internet

link.springer.com
15 <1%
Internet

RMIT University on 2023-05-28


16 <1%
Submitted works

University of Sydney on 2022-10-21


17 <1%
Submitted works

arxiv.org
18 <1%
Internet

data.mendeley.com
19 <1%
Internet

University at Buffalo on 2023-11-25


20 <1%
Submitted works

Sources overview
Similarity Report ID: oid:27005:52404602

University of East London on 2023-05-09


21 <1%
Submitted works

A. Lakshmi, R. Devi. "Heart Disease Prediction Model using Genetic Alg...


22 <1%
Crossref

Ton Duc Thang University


23 <1%
Publication

dokumen.pub
24 <1%
Internet

ejmcm.com
25 <1%
Internet

"Soft Computing and Signal Processing", Springer Science and Busines...


26 <1%
Crossref

Aalto Yliopisto on 2018-12-06


27 <1%
Submitted works

Anisha P Rodrigues, Roshan Fernandes, Aakash A, Abhishek B, Adarsh...


28 <1%
Crossref

Universiti Teknologi MARA on 2020-07-14


29 <1%
Submitted works

University of East London on 2023-09-10


30 <1%
Submitted works

University of Essex on 2023-03-24


31 <1%
Submitted works

mdpi-res.com
32 <1%
Internet

Sources overview
Similarity Report ID: oid:27005:52404602

thesai.org
33 <1%
Internet

"Smart Health", Springer Science and Business Media LLC, 2019


34 <1%
Crossref

Atta Ur Rahman, Yousef Alsenani, Adeel Zafar, Kalim Ullah, Khaled Rab...
35 <1%
Crossref

Central Queensland University on 2023-01-31


36 <1%
Submitted works

University College London on 2023-04-17


37 <1%
Submitted works

Siksha 'O' Anusandhan University on 2023-06-07


38 <1%
Submitted works

Southampton Solent University on 2024-01-08


39 <1%
Submitted works

Sources overview

You might also like