Volume 23, Issue 1 (5-2023)
ijdld 2023, 23(1): 53-67 |
Back to browse issues page
Department of Industrial Engineering, Bandar Abbas Branch, Islamic Azad University, Bandar Abbas, Iran , n.rafiei@iau-tnb.ac.ir
Abstract: (692 Views)
Background: Diabetes entails a great quantity of deaths each year and a great quantity of people living with the disease do not find out their health status early sufficient. In this paper, we advance a data mining-based model for prematurely diagnosis and prediction of diabetes.
Methods: Although K-means is simple and can be utilized for a vast diversity of data kinds, it is wholly sensitive to initial locations of cluster centers which specify the final cluster result, which either enables an efficiently and adequate clustered dataset for the logistic regression model, or presents a lesser amount of data as a result of wrong clustering of the main dataset, thereby restricting the proficiency of the logistic regression model. The main purpose of this study is was to specify procedures of ameliorating the k-means clustering and logistic regression accuracy consequence. Therefore, our algorithm comprises of principal component analysis technique, k-means technique and logistic regression model.
Results: The results obtained from this study show that the ability to obtain the result of K-means clustering accuracy is much higher than what other researchers have obtained in similar studies. Also, compared to the results obtained from other algorithms, the logistic regression model was implemented at an improved level in predicting the onset of diabetes. Another real advantage is that the proposed algorithm was able to successfully model a new dataset.
Conclusion: In general, the proposed approach can be effectively used in predicting and early diagnosis of diabetes.
Methods: Although K-means is simple and can be utilized for a vast diversity of data kinds, it is wholly sensitive to initial locations of cluster centers which specify the final cluster result, which either enables an efficiently and adequate clustered dataset for the logistic regression model, or presents a lesser amount of data as a result of wrong clustering of the main dataset, thereby restricting the proficiency of the logistic regression model. The main purpose of this study is was to specify procedures of ameliorating the k-means clustering and logistic regression accuracy consequence. Therefore, our algorithm comprises of principal component analysis technique, k-means technique and logistic regression model.
Results: The results obtained from this study show that the ability to obtain the result of K-means clustering accuracy is much higher than what other researchers have obtained in similar studies. Also, compared to the results obtained from other algorithms, the logistic regression model was implemented at an improved level in predicting the onset of diabetes. Another real advantage is that the proposed algorithm was able to successfully model a new dataset.
Conclusion: In general, the proposed approach can be effectively used in predicting and early diagnosis of diabetes.
Type of Study: Applicable |
Subject:
General
Received: 2022/12/19 | Accepted: 2023/02/22 | Published: 2023/05/31
Received: 2022/12/19 | Accepted: 2023/02/22 | Published: 2023/05/31
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |