Volume 7, Issue 3 (12-2020)
jhbmi 2020, 7(3): 252-262 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Parsapour F, Peymani J. Using Data Mining Techniques for Intelligent Diagnosis of Severity of Depressive Disorder. jhbmi 2020; 7 (3) :252-262
URL: http://jhbmi.ir/article-1-418-en.html
URL: http://jhbmi.ir/article-1-418-en.html
Ph.D. in Neuroscience, Assistant Professor, Clinical Psychology Dept., Karaj Branch, Islamic Azad University, Alborz, Iran
Abstract: (2686 Views)
Introduction: Implementing a method that can help individuals diagnose or prevent mental disorders can be an important step in preventing and controlling these disorders especially in the early stages. The objective of this research was to apply data mining techniques for intelligent diagnosis of severity of depressive disorder.
Method: The present applied research was carried out by going to a number of psychiatric clinics in Tehran and investigating patients' medical records. A total of 420 subjects who responded to the Minnesota Multiphasic Personality Inventory (MMPI, 71 questions) were selected through convenience sampling as the sample of the study (300 subjects were diagnosed with different degrees of depression and 120 subjects showed no symptoms of depression). The answer sheet of MMPI and the diagnosis of the psychiatrist were used as data for developing the model by k--Nearest Neighbor (k-NN), Decision Tree, and Support Vector Machine algorithms. About 70 percent of the data were applied for training and 30 percent of the data were used for validating the model. MATLAB software was used for data analysis.
Results: The results of the evaluations showed that Decision Tree algorithm with accuracy of 99.16% had higher accuracy compared to other algorithms. Furthermore, by implementing the developed models on each question of MMPI, the influence of each question on evaluation was determined.
Conclusion: Classifying patients with data mining approach and based on the most important characteristics can be a useful and effective tool for analyzing and improving the decision-making process of physicians regarding the treatment of patients.
Method: The present applied research was carried out by going to a number of psychiatric clinics in Tehran and investigating patients' medical records. A total of 420 subjects who responded to the Minnesota Multiphasic Personality Inventory (MMPI, 71 questions) were selected through convenience sampling as the sample of the study (300 subjects were diagnosed with different degrees of depression and 120 subjects showed no symptoms of depression). The answer sheet of MMPI and the diagnosis of the psychiatrist were used as data for developing the model by k--Nearest Neighbor (k-NN), Decision Tree, and Support Vector Machine algorithms. About 70 percent of the data were applied for training and 30 percent of the data were used for validating the model. MATLAB software was used for data analysis.
Results: The results of the evaluations showed that Decision Tree algorithm with accuracy of 99.16% had higher accuracy compared to other algorithms. Furthermore, by implementing the developed models on each question of MMPI, the influence of each question on evaluation was determined.
Conclusion: Classifying patients with data mining approach and based on the most important characteristics can be a useful and effective tool for analyzing and improving the decision-making process of physicians regarding the treatment of patients.
Type of Study: Original Article |
Subject:
Artificial Intelligence in Healthcare
Received: 2019/07/14 | Accepted: 2020/02/3
Received: 2019/07/14 | Accepted: 2020/02/3
Audio file [M4A 3507 KB] (86 Download)
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
