A Comparative Study of Data Mining Clustering Methods for Diabetes Prediction in Healthcare Data Set

Abstract

Diabetes is a disease which is affecting many people now-a-days. Diabetes is a chronic disease caused due to the expanded level of sugar addiction in the blood. Various automated information systems were outlined utilizing various classifiers for anticipate and diagnose the diabetes. Due to its continuously increasing rate, more and more families are unfair by diabetes mellitus. Most diabetics know little about their risk factor they face prior to diagnosis. Data mining approach helps to diagnose patient’s diseases. It has played an important role in diabetes research. It would be a valuable asset for diabetes researchers because it can unearth hidden knowledge from a huge amount of diabetes-related data. Various data mining techniques help diabetes research and ultimately improve the quality of health care for diabetes patients. The primary target of this examination is to assemble Intelligent Diabetes Disease Prediction System that gives analysis of diabetes malady utilizing diabetes patient\'s database .Clustering is the process of partitioning the data or objects into the same class and data in one class is more similar to each other than to those in other cluster. The process of partitioning data objects into subclasses is called as cluster. The quality of cluster depends on the method used. In this research we present the comparison of different clustering techniques using Waikato Environment for Knowledge Analysis or in short (WEKA) by using diabetes data set. The algorithm or methods tested are Density Based Clustering, filtered Cluster and K-MEANS clustering algorithms. This research present a comparative analysis for various clustering techniques on diabetes datasets.

Introduction

This study focuses on the application of data mining techniques for diabetes prediction, with particular emphasis on comparing different clustering algorithms using the WEKA data mining tool. Data mining is the process of extracting meaningful patterns and knowledge from large datasets using techniques such as classification, clustering, association, prediction, regression, and outlier detection. Classification assigns data to predefined categories using methods such as Decision Trees, Bayesian classifiers, Neural Networks, and Genetic Algorithms, while clustering groups similar data objects without predefined classes.

The data mining process consists of several sequential stages:

1. Data Cleaning – Removing incomplete, noisy, inconsistent, or erroneous data.
2. Data Integration – Combining data from multiple sources while reducing redundancy.
3. Data Selection – Selecting relevant data for analysis.
4. Data Transformation – Converting data into suitable formats through normalization and aggregation.
5. Data Mining – Applying algorithms to discover patterns and relationships.
6. Pattern Evaluation – Identifying useful and meaningful patterns.
7. Knowledge Representation – Presenting the extracted knowledge in an understandable form.

The study also discusses major data mining techniques, including:

• Association for discovering relationships between variables.
• Clustering for grouping similar patients.
• Classification for assigning records to predefined classes.
• Prediction for forecasting outcomes.
• Outlier Detection for identifying unusual patterns.
• Regression for modeling relationships between variables.
• Pattern Tracking for recognizing trends over time.
• Evolution and Deviation Analysis for analyzing time-dependent changes.

The proposed research methodology compares the performance of three clustering algorithms:

• K-Means
• DBSCAN (Make Density-Based Clusterer)
• Filtered Cluster

The methodology involves selecting the algorithms, loading a diabetes dataset in CSV format into WEKA, normalizing the data, applying each clustering algorithm to both normalized and unnormalized datasets, recording the results, and comparing their performance using graphical analysis.

The diabetes dataset used contains:

• 768 instances
• 9 attributes
• Multivariate data related to diabetes diagnosis.

Performance evaluation is based on two primary metrics:

• Accuracy, which measures correctly classified instances.
• Execution Time, which measures the time required to build the clustering model.

Experimental results show that:

• DBSCAN achieved the lowest Sum of Squared Error (10,197) compared to K-Means (14,951) and Filtered Cluster (26,729), indicating better clustering quality.
• DBSCAN also had the fastest execution time (0.01 seconds), followed by Filtered Cluster (0.02 seconds) and K-Means (0.05 seconds).

Conclusion

A chronic disease that causes major causalities in the worldwide that is Diabetes. As per International Diabetes Federation (IDF) around the world estimated 285 million people are suffering from diabetes. This range and data will increase in nearby future as there is no appropriate method till date that minimize the effects and prevent it completely. Type 2 diabetes (TTD) is the most common type of diabetes. The major issue was the detection of TTD as it was not easy to predict all the effects. Therefore, data mining was used as it provides the optimal results and help in knowledge discovery from data. In this work, it is concluded that diabetes prediction is the complex problem due to high complexity of the dataset. In this research shows comparative study has been performed on the K- means, filtered cluster and Density based clustering algorithms. Comparison is performed on Bank and segmentation dataset using WEKA tool and the comparative results are presented in the form of table and graph. The comparative study is performed on the basis of accuracy and efficiency parameters. Every algorithm has their own significance and we use them on the nature of the data, but on the basis of this research we concluded that k-means clustering algorithm is simplest algorithm as compared to other algorithms. Filtered cluster is more accurate than other clustering algorithms in case of error rate but in case of execution time the DBSCAN is better than other clustering algorithms. Clustering is a vivid method. The solution is not exclusive and it firmly depends upon the analysts’ choices. Clustering always provides groups or clusters, even if there is no predefined structure. While applying cluster analysis we are contemplating that the groups exist. But this speculation may be false. The outcome of clustering should never be generalized.

References

[1] Ioannis Kavakiotis, Olga Tsave, AthanasiosSalifoglou, NicosMaglaveras, IoannisVlahavas, IoannaChouvarda, “Machine Learning and Data Mining Methods in Diabetes Research”, Computational and Structural Biotechnology Journal 15 (2017) 104–116 [2] BayuAdhi Tama,1 Afriyan Firdaus,2 Rodiyatul FS, “Detection of Type 2 Diabetes Mellitus with Data Mining Approach Using Support Vector Machine”, Vol. 11, issue 3, pp. 12-23, 2008. [3] Yu-Xuan Wang, QiHui Sun, Ting-Ying Chien, Po-Chun Huang, “Using Data Mining and Machine Learning Techniques for System Design Space Exploration and Automatized Optimization”, Proceedings of the 2017 IEEE International Conference on Applied System Innovation, vol. 15, pp. 1079-1082, 2017. [4] ZhiqiangGe, Zhihuan Song, Steven X. Ding, Biao Huang, “Data Mining and Analytics in the Process Industry: The Role of Machine Learning”, 2017 IEEE. Translations and content mining are permitted for academic research only, vol. 5, pp. 20590-20616, 2017. [5] JahinMajumdar, Anwesha Mal, Shruti Gupta, “Heuristic Model to Improve Feature Selection Based on Machine Learning in Data Mining”, 2016 6th International Conference - Cloud System and Big Data Engineering (Confluence), vol. 3, pp. 73-77, 2016. [6] MS.Tejashri n. Giri, prof. S.r.todamal, “data mining approach for diagnosing type 2 diabetes”, international journal of science, engineering and technology, vol. 2 issue 8, 2014. [7] P. Suresh Kumar and V. Umatejaswi, “ Diagnosing Diabetes using Data Mining Techniques”, International Journal of Scientific and Research Publications, Volume 7, Issue 6, June 2017. [8] M. Sharma, G. Singh, R. Singh, “Stark Assessment of Lifestyle Based Human Disorders Using Data Mining Based Learning Techniques”, Elsevier, vol. 5, pp. 202-222, 2017. [9] Han Wu, Shengqi Yang, Zhangqin Huang, Jian He, Xiaoyi Wang, “Type 2 diabetes mellitus prediction model based on data mining”, ScienceDirect, Vol. 11, issue 3, pp. 12-23, 2018. [10] Yan Luo, Charles Ling, Ph.D., Jody Schuurman, Robert Petrella, MD, “GlucoGuide: An Intelligent Type-2 Diabetes Solution Using Data Mining and Mobile Computing”, 2014 IEEE International Conference on Data Mining, Vol. 9, issue 8, pp. 12-23, 2014. [11] AiswaryaIyer, S. Jeyalatha and RonakSumbaly, “DIAGNOSIS OF DIABETES USING CLASSIFICATION MINING TECHNIQUES”, International Journal of Data Mining & Knowledge Management Process (IJDKP) Vol.5, No.1, 2015. [12] Alexis Marcano-Cede˜no, Diego Andina, “Data mining for the diagnosis of type 2 diabetes”, IEEE, Vol. 11, issue 3, pp. 9-19, 2016. [13] B. M. Patil, R. C. Joshi, DurgaToshniwal, “Association rule for classification of type -2 diabetic patients”, 2010 Second International Conference on Machine Learning and Computing, Vol. 8, issue 3, pp. 7-23, 2010. [14] Prova Biswas1,2, Ashoke Sutradhar3, PallabDatta, “Estimation of parameters for plasma glucose regulation in type-2 diabetics in presence of meal”, IET Syst. Biol., 2018, Vol. 12 Iss. 1, pp. 18-25, 2018. [15] Debadri Dutta, Debpriyo Paul, Parthajeet Ghosh, “Analysing Feature Importances for Diabetes Prediction using Machine Learning”, 2018 IEEE 9th Annual Information Technology, Electronics and Mobile Communication Conference (IEMCON) [16] AyushAnand, Divya Shakti, “Prediction of diabetes based on personal lifestyle indicators”, 2015 1st International Conference on Next Generation Computing Technologies (NGCT), Conference Paper [17] SunghwanBae, Taesung Park, “Risk prediction using common and rare genetic variants: Application to Type 2 diabetes”, 2017 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) [18] Sneha Joshi, MeghaBorse, “Detection and Prediction of Diabetes Mellitus Using Back-Propagation Neural Network”, 2016 International Conference on Micro-Electronics and Telecommunication Engineering (ICMETE) [19] MessanKomi, Jun Li, YongxinZhai, Xianguo Zhang, “Application of data mining methods in diabetes prediction”, 2017 2nd International Conference on Image, Vision and Computing (ICIVC) [20] SalimChemlal, Sheri Colberg, Marta Satin-Smith, Eric Gyuricsko, Tom Hubbard, Mark W. Scerbo, Frederic D. McKenzie, “Blood glucose individualized prediction for type 2 diabetes using iPhone application”, 2011 IEEE 37th Annual Northeast Bioengineering Conference (NEBEC)

Copyright

Copyright © 2026 Kamalpreet Kaur, Er. Bharti, Dr. Naveen Dhillion. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.