Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


ucin1097000476.pdf (1.04 MB)

ETD Abstract Container

Abstract Header

COMPARISON OF NEURAL NETWORK AND LOGISTIC REGRESSION MODELS TO PREDICT MEDICAL OUTCOME

VENKATARAMAN, AARTI
http://rave.ohiolink.edu/etdc/view?acc_num=ucin1097000476

Abstract Details

2004, MS, University of Cincinnati, Engineering : Industrial Engineering.
More reliable prediction of outcome would be helpful for clinicians who treat patients with urinary incontinence. A model, which can predict the outcome of stress urinary incontinence surgery, could aid definition of subgroups of patients who might benefit most or least from stress urinary incontinence surgery and therefore guide surgery in terms of uptake or avoidance. The purpose of this research was to develop a neural network (NN) model and a logistic regression model for predicting the Incontinence Symptoms Severity of patients who have undergone surgery for stress urinary incontinence and to compare the results of both the models developed. The data for this research was obtained from a study conducted at Departme nt of Urology, University of Michigan. The artificial neural network was trained using 225 clinical sets using error back propagation and validated through independent testing of 200 records. Eighteen inputs were used to predict categorical output value. Logistic regression analysis was performed using the same development and validation datasets to provide a comparison. The neural network model had a higher sensitivity (78.3% versus 62.5%), specificity (75% versus 54.2%), area under the ROC curve (0.74 versus 0.72) and less error rate (22.5% versus 39.5%) compared to the logistic regression model. The predictive values and the likelihood ratios test both indicated better performance of neural networks over logistic regression analysis. To conclude, neural network could provide a useful predictive model for the optimization of limited resources. The neural network is a new alternative classifying method for developing a predictive paradigm, and it has a higher classifying performance compared to the logistic regression model.
Dr. Ernest Hall (Advisor)
86 p.
Engineering, Industrial
Neural Networks; Logistic Regression; Predict Outcome

Recommended Citations

Citations

  • VENKATARAMAN, A. (2004). COMPARISON OF NEURAL NETWORK AND LOGISTIC REGRESSION MODELS TO PREDICT MEDICAL OUTCOME [Master's thesis, University of Cincinnati]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1097000476

    APA Style (7th edition)

  • VENKATARAMAN, AARTI. COMPARISON OF NEURAL NETWORK AND LOGISTIC REGRESSION MODELS TO PREDICT MEDICAL OUTCOME. 2004. University of Cincinnati, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=ucin1097000476.

    MLA Style (8th edition)

  • VENKATARAMAN, AARTI. "COMPARISON OF NEURAL NETWORK AND LOGISTIC REGRESSION MODELS TO PREDICT MEDICAL OUTCOME." Master's thesis, University of Cincinnati, 2004. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1097000476

    Chicago Manual of Style (17th edition)

ucin1097000476
897
© 2004, all rights reserved.
This open access ETD is published by University of Cincinnati and OhioLINK.