Sensitivity Analysis by Variables in the Classification of Renal Insufficiency Using Artificial Neural Networks VS Multinomial Logistic Regression

Elena Martín Pérez; Quintín Martín Martín

doi:doi:10.11648/j.bsi.20220701.12

| Peer-Reviewed

Sensitivity Analysis by Variables in the Classification of Renal Insufficiency Using Artificial Neural Networks VS Multinomial Logistic Regression

Elena Martín Pérez, Quintín Martín Martín

Published in Biomedical Statistics and Informatics (Volume 7, Issue 1)

Received: 18 February 2022 Accepted: 14 March 2022 Published: 23 March 2022

Views: Downloads:

Download PDF

Share This Article

Twitter
Linked In
Facebook

Abstract

Purpose: The aim of the research is to see to what extent Neural Networks (non-parametric technique) could be used as opposed to Multinomial Logistic Regression (parametric technique) in the analysis of sensitivity by variable in the classification of "Renal insufficiency" (three categories: NOT, MODERATE and ADVANCED). The analysis of sensitivity by variable, in our case, consists of eliminating from the model the three most influential variables (one by one): Blood creatine, Urine creatine and Urea, in that order. Once a variable is removed from the model, it will not reenter the model. Methods: This study collects data from the University Hospital of Salamanca (Spain), configuring a file of renal insufficiency data with 184 cases and 9 variables. First, we do descriptive-exploratory analysis of data for renal insufficiency data, obtained experimentally through a pilot survey. The comparison between ANNs and MLR is carried out by classification in the categories NOT, MODERATE and ADVANCED. Results: The descriptive-exploratory analysis of data shows the high value of the Coefficient of variation, Kurtosis and Skewness of the variables Blood creatine, Urine creatine and Urea, all of them well above 66%, 10 and 2.70, respectively. The study shows that when all the variables in the model are considered, the highest classification percentage concerns the Multinomial Logistic Regression (MLR), while, for the analysis of sensitivity by variables, the classification percentages are favourable to the Artificial Neural Network model (ANN). Conclusions: The joint classification percentages in the analysis of sensitivity by variables are favourable to the artificial neural network model (perceptron). That is, the non-parametric technique (ANNs) would surpass the parametric technique (MLR) in the classification of a patient's renal insufficiency.

Published in	Biomedical Statistics and Informatics (Volume 7, Issue 1)
DOI	10.11648/j.bsi.20220701.12
Page(s)	7-11
Creative Commons	This is an Open Access article, distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium or format, provided the original work is properly cited.
Copyright	Copyright © The Author(s), 2022. Published by Science Publishing Group

Keywords

Renal Insufficiency, Exploratory Data Analysis, Artificial Neural Network, Multinomial Logistic Regression

References

[1]	Tu, J. V. (1996). Advantages and Disadvantages of Using Artificial Neural Networks versus Logistic Regression for Predicting Medical Outcomes. Journal of Clinical Epidemiology, 49 (11), 1225-1231.
[2]	Lang, E. W., Pitts, L. H., Damron, S. L. & Rutledge, R. (1997). Outcome after Severe Head Injury: An Analysis of Prediction Based upon Comparison of Neural Network versus Logistic Regression Analysis. Neurol Res, 19, 274-80.
[3]	Terrin, N., Schmid, C. H., Griffith, J. L., D'Agostino, R. B. & Selker, H. P. (2003). External Validity of Predictive Models: A Comparison of Logistic Regression, Classification Trees, and Neural Networks. Journal of Clinical Epidemiology. 56, 721-729.
[4]	Song, J. H, Venkatesh, S. S., Conant, E. A., Arger, P. H. & Sehgal, C. M. (2005). Comparative Analysis of Logistic Regression and Artificial Neural Network for Computer-aided Diagnosis of Breast Masses. Acad Radiol, 12 (4), 487-95.
[5]	Chen, S. T., Hsiao, Y. H., Huang, Y. L., Kuo, S. J., Tseng, H. S., Wu, H. K. et al. (2009). Comparative Analysis of Logistic Regression, Support Vector Machine and Artificial Neural Network for the Differential Diagnosis of Benign and Malignant Solid Breast Tumors by the Use of Three-dimensional Power Doppler Imaging. Korean Journal of Radiology, 10 (5), 464-471.
[6]	Shafiei, E. et al. (2017). Comparison of Artificial Neural Network and Logistic Regression Models for Prediction of Psychological Symptom Six Months after Mild Traumatic Brain Injury. Iran J Psychiatry Behav Sci.
[7]	Martín Pérez, E., Caldero Alonso, A. & Martín Martín, Q. (2021). Classification of Psychiatric Disorders Using Multinomial Logistic Regression vs Artificial Neural Network. JP Journal of Biostatistics, 18, (3), 395-408.
[8]	Ayer, T., Chatwal, J., Alagoz, O., Kahn, C. E., Jr., Woods, R. W. & Burnside, E. S. (2010). Informatics in Radiology: Comparison of Logistic Regression and Artificial Neural Network Models in Breast Cancer Risk Estimation. Radiographics, 30 (1), 13-22.
[9]	Bartfay, E., Mackillop, W. J. & Pater, J. L. (2006). Comparing the Predictive Value of Neural Network Models to Logistic Regression Models on the Risk of Death for Small-cell Lung Cancer Patients. European Journal of Cancer Care (England), 15 (2), 115-124.
[10]	Caocci, G., Baccoli, R., Vacca, A., Mastronuzzi, A., Bertaina, A., Piras, E. et al. (2010). Comparison between an Artificial Neural Network and Logistic Regression in Predicting Acute Graft-vs-Host Disease after Unrelated Donor Hematopoietic Stem Cell Transplantation in Thalassemia Patients. Experimental Hematology, 38 (5), 426-433.
[11]	Chen, H., Zhang, J., Xu, Y., Chen, B. & Zhang, K. (2012). Performance Comparison of Artificial Neural Network and Logistic Regression Model for Differentiating Lung Nodules on CT Scans. Expert Systems with Applications, 39, 11503-11509.
[12]	Dreiseitl, S. & Ohno-Machado, L. (2002). Logistic Regression and Artificial Neural Network Classification Models: A Methodology Review. Journal of Biomedical Informatics, 35 (5-6), 352-359.
[13]	Jaimes, F., Farbiarz, J., Álvarez, D. & Martínez, C. (2005). Comparison between Logistic Regression and Neural Networks to Predict Death in Patients with Suspected Sepsis in the Emergency Room. Critical Care, 9 (2), R150-156.
[14]	Kazemnejad, A., Batvandi, Z. & Faradmal, J. (2010). Comparison of Artificial Neural Network and Binary Logistic Regression for Determination of Impaired Glucose Tolerance/Diabetes. Eastern Mediterranean Health Journal, 16 (6), 615-620.
[15]	Lin, C. C., Ou, Y. K., Chen, S. H., Liu, Y. C. & Lin, J. (2010). Comparison of Artificial Neural Network and Logistic Regression Models for Predicting Mortality in Elderly Patients with Hip Fracture. Injury, 41 (8), 869-873.
[16]	Lin, S. P., Lee, C. H., Lu, Y. S. & Hsu, L. N. (2006). A Comparison of MICU Survival Prediction Using the Logistic Regression Model and Artificial Neural Network Model. Journal of Nursing Research, 14 (4), 306-314.
[17]	Ottenbacher et al. (2004). Comparison of Logistic Regression and Neural Network Analysis Applied to Predicting Living Setting after Hip Fracture. Annals of Epidemiology, 14 (8), 551-559.
[18]	Pai, D. R. et al. (2012). Experimental Comparison of Parametric, Non-parametric, and Hybrid Multigroup Classification. Expert Systems with Applications, 39 (10), 8593-8603.
[19]	Sargent, D. J. (2001). Comparison of artificial neural networks with other statistical approaches: Results from medical data sets. Cancer, 91 (sup 8), 1636–1642.
[20]	Masic, N. & Pfurtscheller, G. (1993). Neural Network Based Classification of Single-trial EEG Data. Artificial Intelligence in Medicine, 5, 1993, 503-513.
[21]	Dorffner, G. & Porenta, G. (1994). On Using Feedforward Neural Networks for Clinical Diagnostic Tasks. Artificial Intelligence in Medicine, 6, 417-435.
[22]	Dybowski, R. & Gant, V. (2001). Clinical Applications of Artificial Neural Networks. Cambridge University Press.
[23]	Mazurowski, M. A. et al. (2008). Training Neural Network Classifiers for Medical Decision Making: The Effects of Imbalanced Datasets on Classification Performance. International Joint Conference on Neural Networks IJCNN’07. Orlando, Florida, USA, 21 (3), 427-436.
[24]	Ince, T. et al. (2010). Evaluation of Global and Local Training Techniques over Feed-forward Neural Network Architecture Spaces for Computer-aided Medical Diagnosis. Expert Systems with Applications, 37, 8450-8461.
[25]	Amato F. et al. (2013). Artificial Neural Networks in Medical Diagnosis. Journal of Applied Biomedicine, 11, 47-58.
[26]	Peek, N. et al. (2015). Thirty Years of Artificial Intelligence in Medicine (AIME) Conferences: A Review of Research Themes. Artificial Intelligence in Medicine, 65, 61-73.
[27]	Shaikhina, T. & Khovanova, N. A. (2017). Handling Limited Datasets with Neural Networks in Medical Applications: A Small-data Approach. Artificial Intelligence in Medicine, 75, 51-63.
[28]	Corsinia, M.-M., Schmittb, A. & Bruzekb, J. (2005). Process Variability on the Human Skeleton: Artificial Network as an Appropriate Tool for Age at Death Assessment. Forensic Science International, 148, 163-167.
[29]	Barni, M., Faila, P. & Lazzareti, R. Privacy-preserving ECG (2011). Classification with Branching Programs and Neural Networks. IEEE Transactions on Information Forensics and Security, 6 (2), 452-468.

Cite This Article

Plain Text BibTeX RIS

APA Style

Elena Martín Pérez, Quintín Martín Martín. (2022). Sensitivity Analysis by Variables in the Classification of Renal Insufficiency Using Artificial Neural Networks VS Multinomial Logistic Regression. Biomedical Statistics and Informatics, 7(1), 7-11. https://doi.org/10.11648/j.bsi.20220701.12

Copy | Download

ACS Style

Elena Martín Pérez; Quintín Martín Martín. Sensitivity Analysis by Variables in the Classification of Renal Insufficiency Using Artificial Neural Networks VS Multinomial Logistic Regression. Biomed. Stat. Inform. 2022, 7(1), 7-11. doi: 10.11648/j.bsi.20220701.12

Copy | Download

AMA Style

Elena Martín Pérez, Quintín Martín Martín. Sensitivity Analysis by Variables in the Classification of Renal Insufficiency Using Artificial Neural Networks VS Multinomial Logistic Regression. Biomed Stat Inform. 2022;7(1):7-11. doi: 10.11648/j.bsi.20220701.12

Copy | Download

@article{10.11648/j.bsi.20220701.12,
  author = {Elena Martín Pérez and Quintín Martín Martín},
  title = {Sensitivity Analysis by Variables in the Classification of Renal Insufficiency Using Artificial Neural Networks VS Multinomial Logistic Regression},
  journal = {Biomedical Statistics and Informatics},
  volume = {7},
  number = {1},
  pages = {7-11},
  doi = {10.11648/j.bsi.20220701.12},
  url = {https://doi.org/10.11648/j.bsi.20220701.12},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.bsi.20220701.12},
  abstract = {Purpose: The aim of the research is to see to what extent Neural Networks (non-parametric technique) could be used as opposed to Multinomial Logistic Regression (parametric technique) in the analysis of sensitivity by variable in the classification of "Renal insufficiency" (three categories: NOT, MODERATE and ADVANCED). The analysis of sensitivity by variable, in our case, consists of eliminating from the model the three most influential variables (one by one): Blood creatine, Urine creatine and Urea, in that order. Once a variable is removed from the model, it will not reenter the model. Methods: This study collects data from the University Hospital of Salamanca (Spain), configuring a file of renal insufficiency data with 184 cases and 9 variables. First, we do descriptive-exploratory analysis of data for renal insufficiency data, obtained experimentally through a pilot survey. The comparison between ANNs and MLR is carried out by classification in the categories NOT, MODERATE and ADVANCED. Results: The descriptive-exploratory analysis of data shows the high value of the Coefficient of variation, Kurtosis and Skewness of the variables Blood creatine, Urine creatine and Urea, all of them well above 66%, 10 and 2.70, respectively. The study shows that when all the variables in the model are considered, the highest classification percentage concerns the Multinomial Logistic Regression (MLR), while, for the analysis of sensitivity by variables, the classification percentages are favourable to the Artificial Neural Network model (ANN). Conclusions: The joint classification percentages in the analysis of sensitivity by variables are favourable to the artificial neural network model (perceptron). That is, the non-parametric technique (ANNs) would surpass the parametric technique (MLR) in the classification of a patient's renal insufficiency.},
 year = {2022}
}

Copy | Download

TY  - JOUR
T1  - Sensitivity Analysis by Variables in the Classification of Renal Insufficiency Using Artificial Neural Networks VS Multinomial Logistic Regression
AU  - Elena Martín Pérez
AU  - Quintín Martín Martín
Y1  - 2022/03/23
PY  - 2022
N1  - https://doi.org/10.11648/j.bsi.20220701.12
DO  - 10.11648/j.bsi.20220701.12
T2  - Biomedical Statistics and Informatics
JF  - Biomedical Statistics and Informatics
JO  - Biomedical Statistics and Informatics
SP  - 7
EP  - 11
PB  - Science Publishing Group
SN  - 2578-8728
UR  - https://doi.org/10.11648/j.bsi.20220701.12
AB  - Purpose: The aim of the research is to see to what extent Neural Networks (non-parametric technique) could be used as opposed to Multinomial Logistic Regression (parametric technique) in the analysis of sensitivity by variable in the classification of "Renal insufficiency" (three categories: NOT, MODERATE and ADVANCED). The analysis of sensitivity by variable, in our case, consists of eliminating from the model the three most influential variables (one by one): Blood creatine, Urine creatine and Urea, in that order. Once a variable is removed from the model, it will not reenter the model. Methods: This study collects data from the University Hospital of Salamanca (Spain), configuring a file of renal insufficiency data with 184 cases and 9 variables. First, we do descriptive-exploratory analysis of data for renal insufficiency data, obtained experimentally through a pilot survey. The comparison between ANNs and MLR is carried out by classification in the categories NOT, MODERATE and ADVANCED. Results: The descriptive-exploratory analysis of data shows the high value of the Coefficient of variation, Kurtosis and Skewness of the variables Blood creatine, Urine creatine and Urea, all of them well above 66%, 10 and 2.70, respectively. The study shows that when all the variables in the model are considered, the highest classification percentage concerns the Multinomial Logistic Regression (MLR), while, for the analysis of sensitivity by variables, the classification percentages are favourable to the Artificial Neural Network model (ANN). Conclusions: The joint classification percentages in the analysis of sensitivity by variables are favourable to the artificial neural network model (perceptron). That is, the non-parametric technique (ANNs) would surpass the parametric technique (MLR) in the classification of a patient's renal insufficiency.
VL  - 7
IS  - 1
ER  -

Copy | Download

Author Information

Elena Martín Pérez

Institute of Legal Medicine of Zamora, Zamora, Spain
Quintín Martín Martín

Department of Statistics, University of Salamanca, Salamanca, Spain

Download PDF

Sections

Plain Text BibTeX RIS

APA Style

Elena Martín Pérez, Quintín Martín Martín. (2022). Sensitivity Analysis by Variables in the Classification of Renal Insufficiency Using Artificial Neural Networks VS Multinomial Logistic Regression. Biomedical Statistics and Informatics, 7(1), 7-11. https://doi.org/10.11648/j.bsi.20220701.12

Copy | Download

ACS Style

Elena Martín Pérez; Quintín Martín Martín. Sensitivity Analysis by Variables in the Classification of Renal Insufficiency Using Artificial Neural Networks VS Multinomial Logistic Regression. Biomed. Stat. Inform. 2022, 7(1), 7-11. doi: 10.11648/j.bsi.20220701.12

Copy | Download

AMA Style

Elena Martín Pérez, Quintín Martín Martín. Sensitivity Analysis by Variables in the Classification of Renal Insufficiency Using Artificial Neural Networks VS Multinomial Logistic Regression. Biomed Stat Inform. 2022;7(1):7-11. doi: 10.11648/j.bsi.20220701.12

Copy | Download

@article{10.11648/j.bsi.20220701.12,
  author = {Elena Martín Pérez and Quintín Martín Martín},
  title = {Sensitivity Analysis by Variables in the Classification of Renal Insufficiency Using Artificial Neural Networks VS Multinomial Logistic Regression},
  journal = {Biomedical Statistics and Informatics},
  volume = {7},
  number = {1},
  pages = {7-11},
  doi = {10.11648/j.bsi.20220701.12},
  url = {https://doi.org/10.11648/j.bsi.20220701.12},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.bsi.20220701.12},
  abstract = {Purpose: The aim of the research is to see to what extent Neural Networks (non-parametric technique) could be used as opposed to Multinomial Logistic Regression (parametric technique) in the analysis of sensitivity by variable in the classification of "Renal insufficiency" (three categories: NOT, MODERATE and ADVANCED). The analysis of sensitivity by variable, in our case, consists of eliminating from the model the three most influential variables (one by one): Blood creatine, Urine creatine and Urea, in that order. Once a variable is removed from the model, it will not reenter the model. Methods: This study collects data from the University Hospital of Salamanca (Spain), configuring a file of renal insufficiency data with 184 cases and 9 variables. First, we do descriptive-exploratory analysis of data for renal insufficiency data, obtained experimentally through a pilot survey. The comparison between ANNs and MLR is carried out by classification in the categories NOT, MODERATE and ADVANCED. Results: The descriptive-exploratory analysis of data shows the high value of the Coefficient of variation, Kurtosis and Skewness of the variables Blood creatine, Urine creatine and Urea, all of them well above 66%, 10 and 2.70, respectively. The study shows that when all the variables in the model are considered, the highest classification percentage concerns the Multinomial Logistic Regression (MLR), while, for the analysis of sensitivity by variables, the classification percentages are favourable to the Artificial Neural Network model (ANN). Conclusions: The joint classification percentages in the analysis of sensitivity by variables are favourable to the artificial neural network model (perceptron). That is, the non-parametric technique (ANNs) would surpass the parametric technique (MLR) in the classification of a patient's renal insufficiency.},
 year = {2022}
}

Copy | Download

TY  - JOUR
T1  - Sensitivity Analysis by Variables in the Classification of Renal Insufficiency Using Artificial Neural Networks VS Multinomial Logistic Regression
AU  - Elena Martín Pérez
AU  - Quintín Martín Martín
Y1  - 2022/03/23
PY  - 2022
N1  - https://doi.org/10.11648/j.bsi.20220701.12
DO  - 10.11648/j.bsi.20220701.12
T2  - Biomedical Statistics and Informatics
JF  - Biomedical Statistics and Informatics
JO  - Biomedical Statistics and Informatics
SP  - 7
EP  - 11
PB  - Science Publishing Group
SN  - 2578-8728
UR  - https://doi.org/10.11648/j.bsi.20220701.12
AB  - Purpose: The aim of the research is to see to what extent Neural Networks (non-parametric technique) could be used as opposed to Multinomial Logistic Regression (parametric technique) in the analysis of sensitivity by variable in the classification of "Renal insufficiency" (three categories: NOT, MODERATE and ADVANCED). The analysis of sensitivity by variable, in our case, consists of eliminating from the model the three most influential variables (one by one): Blood creatine, Urine creatine and Urea, in that order. Once a variable is removed from the model, it will not reenter the model. Methods: This study collects data from the University Hospital of Salamanca (Spain), configuring a file of renal insufficiency data with 184 cases and 9 variables. First, we do descriptive-exploratory analysis of data for renal insufficiency data, obtained experimentally through a pilot survey. The comparison between ANNs and MLR is carried out by classification in the categories NOT, MODERATE and ADVANCED. Results: The descriptive-exploratory analysis of data shows the high value of the Coefficient of variation, Kurtosis and Skewness of the variables Blood creatine, Urine creatine and Urea, all of them well above 66%, 10 and 2.70, respectively. The study shows that when all the variables in the model are considered, the highest classification percentage concerns the Multinomial Logistic Regression (MLR), while, for the analysis of sensitivity by variables, the classification percentages are favourable to the Artificial Neural Network model (ANN). Conclusions: The joint classification percentages in the analysis of sensitivity by variables are favourable to the artificial neural network model (perceptron). That is, the non-parametric technique (ANNs) would surpass the parametric technique (MLR) in the classification of a patient's renal insufficiency.
VL  - 7
IS  - 1
ER  -

Copy | Download