Prediction of Diabetes using Support Vector Machine

Ajay Kumar Tiwari, Avadhesh Kumar Dixit, Piyush Rai

IET, Dr. RML Avadh University, Ayodhya

Keyword: Diabetes, SVM, KNN, Random Forest

Abstract: Diabetes is a very common disease in the world. If diabetes is detected in the early stage, it can be cured

easily. Several machine learning techniques are available to predict diabetes in an earlier stage using a data

set. This paper proposes support vector machine based methods to predict diabetes. This paper also provides

the comparative analysis of Naive Bayes, SVM, KNN, Random Forest, Logistic Regression and Decision

Tree to predict diabetes. In this paper the proposed SVM based approach achieved the accuracy 77.08% that

is better in compare to other machine learning based approaches.

1 INTRODUCTION

Diabetes is a coarse disease in our society. In this

type of disease, a person has high blood sugar, either

insulin production inefficient or the body cell do not

return correctly to insulin, or by both reason. In

human body blood sugar level is controlled by

insulin hormone released by pancreas. When due to

any reason secretion of insulin hormone becomes

irregular, blood sugar level also affected. In this way

a person may be affected from diabetes. The patients

affected from diabetes can be cured by regular

exercise, and by adopting healthy lifestyle. To

control blood sugar level some medicine may be

given or insulin may be given explicitly. To know

whether a person is affected from diabetes, some

diagnosis is required. If we came to know about the

disease in early stage, we may prevent this harmful

disease.

2 RELATED WORKS

In literature there are various machine learning

based approaches used for diabetes prediction. He,

B., Shu, K., & Zhang, H. (2019, August) authors has

been proposed a deep learning based method for the

prediction of diabetes. Fitriyani, N. L., Syafrudin,

M., Alfian, G., & Rhee, J. (2019). authors has been

proposed a disease prediction model to provide an

early prediction for T-2 diabetes based on

individual’s risk factors data.Hasan, M. K., Alam,

M. A., Das, D., Hossain, E., & Hasan, M. (2020).

authors has been proposedk-nearest Neighbour,

Decision Trees, Random Forest, Naive Bayes, and

AdaBoost and Multilayer Perceptron (MLP) method

for the prediction of diabetes.Woldemichael, F. G.,

& Menaria, S. (2018, May). authors has been

proposed to predict diabetes using data mining

techniques.Sarwar, M. A., Kamal, N., Hamid, W., &

Shah, M. A. (2018) authors has been proposed a

predictive analysis using different machine learning

methods for the prediction of diabetes. Rout, M., &

Kaur, A. (2020) authors has been proposed to

predict of diabetes diseaseusing various researchof

machine learning Techniques. Shetty, D., Rit, K.,

Shaikh, S., & Patil, N. (2017) authors has been

proposed K-Nearest Neighbour and the Naïve Bayes

for the prediction of diabetes. Alam, T. M., Iqbal,

M. A., Ali, Y., Wahab, A., Ijaz, S., Baig, T. I., &

Abbas, Z. (2019). authors has been proposed to

predict diabetes using various machine learning

algorithms to determine significant attribute

selection for clustering, prediction, andassociation

rule mining for diabetes. Ahmed, T. M. (2016)

authors has been proposed a new predicted model

base on patients HbA1c. Aljumah et.al, (2013)

authors has been proposed SVM based approach by

using the Dataset of disease in Saudi Arabia to

observe obesity and predicts chances of diabetes in a

person. Chen, Peihua, and Chuandi Pan (2018) have

been proposed diabetes prediction model based on

boosting algorithms. They performed non parametric

testing using two algorithms, Adaboost and

Logitboost on test data of 35669 individuals and got

Tiwari, A., Dixit, A. and Rai, P.

Prediction of Diabetes using Support Vector Machine.

DOI: 10.5220/0010563800003161

In Proceedings of the 3rd International Conference on Advanced Computing and Software Engineering (ICACSE 2021), pages 119-122

ISBN: 978-989-758-544-9

 2022 by SCITEPRESS – Science and Technology Publications, Lda. All r ights reser ved

119

area under characteristics curve 0.99. The authors

Mercaldo, Francesco (2017) worked on the concept

of classification. They used Pima Indians dataset and

obtained precision value 0.770 and recall value

0.775.The authors Patil, Bankat M (2010) proposed

a prediction Model which used simple K-means

algorithm and C4.5 algorithm using Pima Indians

diabetes data and achieved an accuracy of

93.5%.The authors Kavakiotis, Ioannis, et al (2017)

have been proposed SVM based model and got the

accuracy of 85%. The authors Kohli, Pahulpreet

Singh (2018) have been proposed logistic regression

on separate datasets of heart, breast cancer and

diabetes and got accuracy of 80.77%.The authors

Perveen, Sajida, et al.(2016) performed

classification technique, decision tree J48 and

achieved Area under ROC is 0.98.The authors

Sisodia, Deepti (2018)used DT, SVM and NB

classification methods on Pima Indians Diabetes

datasets. NB gave accuracy of 76.30%.

3 MATERIAL AND METHODS

Here, in this paper the description of diabetes

dataset, methodology to compare the performance of

various machine learning models has been provided.

3.1 Data Set Description

In this paper the PIMA Indian diabetes dataset

Thomas, Jensia, et al. (2019) was used which was

taken from Kaggle (https://www.kaggle.com/

uciml/pima-indians-diabetes-database). It is made to

predict diabetes in women more than 21 years of

age. It contains eight attributes or input variables and

one output variable. The attributes are as follows:

Pregnancies: It represents number of pregnancies of

a woman. During pregnancy the glucose level of

women may increase which is called gestational

diabetes. If women got pregnant number of times the

gestational diabetes may leads to diabetes mellitus.

Glucose: It represents glucose concentration in

blood. If glucose concentration in blood increases

than a certain value then it may cause diabetes.

Blood Pressure: It represents BP(diastolic in mm

Hg). Higher diastolic blood pressure increases the

risk of diabetes.

BMI: It represents body mass index (weight

(kg)/height (m)^2). It determines the obesity of the

patient. Hence it is an important metric to predict

diabetes.

Skin Thickness: It represents skin thickness (mm).

In case of varying ratio of muscle mass and fat mass

BMI is not adequate parameter to assess obecity

which may lead to diabetes. Hence triceps skinfold

thickness plays an impotant role to predict the

patient may be diabetic or not.

Insulin: It represents serum insulin(mu U/ml). It is 2

hour serum insulin which indicates that how the

body of a person respond on taking food.

Diabetes Pedigree Function: It is a function which

determines the probability of diabetes on the basis of

diabetic family history of a person.

Age: It is generally observed that person having age

greater than 60 years are m ore prone to diabetes

3.2 Support Vector Machine

Support Vector Machine (SVM) is one of the most

popular supervised learning techniques, which is

used for classification as well as regression analysis.

It is used to find a hyper plane in an N-dimensional

space. SVM is based on discrimination. Support

vectors represent data points that are closet to hyper

plane.

4 RESULT AND COMPARATIVE

ANALYSIS

Here, in this paper the result of proposed approach

and comparative analysis of various Machine

Learning Techniques is presented. In this paper the

different machine learning techniques are used and

results for all techniques are different because the

working principle of each technique is different. The

results are evaluated on the basis of accuracy. Here,

by using the K-Nearest Neighbourachieved an

accuracy of 71.48%,the Decision Table obtain

accuracy of 73.04%, Random Forest achieved an

accuracy of 75.13%, Naïve Bayes technique

obtained accuracy of 75.78% ,Logistic Regression

obtained an accuracy of 76.82% andsupport vector

machine obtained better an accuracy of 77.08%. See

(Table-1)and Figure-1.

Table-1. Accuracy for the prediction diabetes using

Machine learning Techniques

Sl. No. Techniques Accuracy

1 K-Nearest Neighbour(KNN) 71.48%

2 Decision Table(DT) 73.04%

3 Random Forest(RF) 75.13%

4 Naïve Bayes 75.78%

5 Logistic Regression(LR) 76.82%

6 Support Vector Machine (SVM) 77.08%

ICACSE 2021 - International Conference on Advanced Computing and Software Engineering

120

Figure 1: Comparison of accuracy of machine learning

Techniques.

5 CONCLUSION

To predict the early stage of diabetes is one of the

most challenging and important task. If diabetes is

detected in an early stage, it can be cured easily.

Machine learning methods have different power in

different data set. Several machine learning

techniques are available to predict diabetes in an

earlier stage using data set. This paper proposed a

support vector machine based methods to predict

diabetes. This paper also provided the comparative

analysis of Naive Bayes, SVM, KNN, Random

Forest, Logistic Regression and Decision Tree to

predict diabetes. In this paper the proposed SVM

based approach achieved the accuracy 77.08% that

is better in compare to other machine learning based

approaches.

REFERENCES

He, B., Shu, K., & Zhang, H. (2019, August). Diabetes

Diagnosis and Treatment Research Based on Machine

Learning. In 2019 IEEE SmartWorld, Ubiquitous

Intelligence &Computing, Advanced & Trusted

Computing, Scalable Computing & Communications,

Cloud & Big Data Computing, Internet of People and

Smart City Innovation

(SmartWorld/SCALCOM/UIC/ATC/CBDCom/IOP/S

CI) (pp. 675-679). IEEE.Hammoudeh, A., Al-Naymat,

G., Ghannam, I., & Obied, N. (2018). Predicting

hospital readmission among diabetics using deep

learning. Procedia Computer Science, 141, 484-489.

Fitriyani, N. L., Syafrudin, M., Alfian, G., & Rhee, J.

(2019). Development of disease prediction model

based on ensemble learning approach for diabetes and

hypertension. IEEE Access, 7, 144777-144789.

Hasan, M. K., Alam, M. A., Das, D., Hossain, E., &

Hasan, M. (2020). Diabetes prediction using

ensembling of different machine learning

classifiers. IEEE Access, 8, 76516-76531.

Woldemichael, F. G., & Menaria, S. (2018, May).

Prediction of diabetes using data mining techniques.

In 2018 2nd International Conference on Trends in

Electronics and Informatics (ICOEI) (pp. 414-418).

IEEE.

Sarwar, M. A., Kamal, N., Hamid, W., & Shah, M. A.

(2018, September). Prediction of diabetes using

machine learning Algorithms in healthcare. In 2018

24th International Conference on Automation and

Computing (ICAC) (pp. 1-6). IEEE.

Rout, M., & Kaur, A. (2020, June). Prediction of Diabetes

Risk based on Machine Learning Techniques. In 2020

International Conference on Intelligent Engineering

and Management (ICIEM) (pp. 246-251). IEEE.

Shetty, D., Rit, K., Shaikh, S., & Patil, N. (2017, March).

Diabetes disease prediction using data mining. In 2017

international conference on innovations in

information, embedded and communication systems

(ICIIECS) (pp. 1-5). IEEE.

Alam, T. M., Iqbal, M. A., Ali, Y., Wahab, A., Ijaz, S.,

Baig, T. I., & Abbas, Z. (2019). A model for early

prediction of diabetes. Informatics in Medicine

Unlocked, 16, 100204.

Ahmed, T. M. (2016). Using data mining to develop

model for classifying diabetic patient control level

based on historical medical records. Journal of

Theoretical and Applied Information

Technology, 87(2), 316.

Aljumah, Abdullah A., Mohammed Gulam Ahamad, and

Mohammad KhubebSiddiqui. "Application of data

mining: Diabetes health care in young and old

patients." Journal of King Saud University-Computer

and Information Sciences 25.2 (2013): 127-136.

Chen, Peihua, and Chuandi Pan. "Diabetes classification

model based on boosting algorithms."BMC

bioinformatics 19.1 (2018): 109.

Mercaldo, Francesco, Vittoria Nardone, and Antonella

Santone. "Diabetes mellitus affected patients

classification and diagnosis through machine learning

techniques." Procedia computer science 112 (2017):

2519-2528.

Patil, Bankat M., Ramesh Chandra Joshi, and

DurgaToshniwal. "Hybrid prediction model for type-2

diabetic patients." Expert systems with applications

37.12 (2010): 8102-8108.

Kavakiotis, Ioannis, et al. "Machine learning and data

mining methods in diabetes research." Computational

and structural biotechnology journal 15 (2017): 104-

116.

Kohli, Pahulpreet Singh, and ShriyaArora. "Application of

Machine Learning in Disease Prediction."2018 4th

International Conference on Computing

Communication and Automation (ICCCA).IEEE,

2018.

Perveen, Sajida, et al. "Performance analysis of data

mining classification techniques to predict diabetes."

Procedia Computer Science 82 (2016): 115-121.

Sisodia, Deepti, and Dilip Singh Sisodia. "Prediction of

diabetes using classification algorithms."Procedia

computer science 132 (2018): 1578-1585.

65,00%

70,00%

75,00%

80,00%

K‐Nearest

Neighb…

Decision

Table…

Random

Forest…

Naïve

Bayes

Logistic

Regressi…

Support

Vector…

123456

Accuracy

Prediction of Diabetes using Support Vector Machine

121

Thomas, Jensia, et al. "Machine Learning Approach For

Diabetes Prediction." International Journal of

Information 8.2 (2019).

https://www.kaggle.com/uciml/pima-indians-diabetes-

database.

ICACSE 2021 - International Conference on Advanced Computing and Software Engineering

122