### Abstract

The purpose of this study was to demonstrate how a form of neural network analysis could be used to perform survival analysis on censored data, and to compare neural network analysis with the most commonly used technique for this type of analysis, Cox regression. In this study computer simulated data sets were used. The underlying rules connecting prognostic information to the hazard of death were defined to allow the construction of data sets with specific realistic properties that could be used to demonstrate situations in which neural network analysis had particular strengths in comparison with Cox regression modeling. Using these simulated data sets neural network analysis could produce successful predictive models, find interactions between variables, and recognize the importance of variables that contributed to the hazard rate as a complex function of the variables value and in situations where the proportionality of hazards assumption was violated. It was also demonstrated that neural network analysis was not a 'black box ', but could lead to useful insights into the roles played by different prognostic variables in determining patient outcome.

Original language | English |
---|---|

Pages (from-to) | 127-138 |

Number of pages | 12 |

Journal | Cancer Letters |

Volume | 77 |

Issue number | 2-3 |

DOIs | |

Publication status | Published - Mar 15 1994 |

### Fingerprint

### Keywords

- Artificial intelligence
- Censored data
- Cox regression
- Neural networks
- Prognostic factors

### ASJC Scopus subject areas

- Cancer Research
- Molecular Biology
- Oncology

### Cite this

*Cancer Letters*,

*77*(2-3), 127-138. https://doi.org/10.1016/0304-3835(94)90095-7

**A technique for using neural network analysis to perform survival analysis of censored data.** / De Laurentiis, Michele; Ravdin, Peter M.

Research output: Contribution to journal › Article

*Cancer Letters*, vol. 77, no. 2-3, pp. 127-138. https://doi.org/10.1016/0304-3835(94)90095-7

}

TY - JOUR

T1 - A technique for using neural network analysis to perform survival analysis of censored data

AU - De Laurentiis, Michele

AU - Ravdin, Peter M.

PY - 1994/3/15

Y1 - 1994/3/15

N2 - The purpose of this study was to demonstrate how a form of neural network analysis could be used to perform survival analysis on censored data, and to compare neural network analysis with the most commonly used technique for this type of analysis, Cox regression. In this study computer simulated data sets were used. The underlying rules connecting prognostic information to the hazard of death were defined to allow the construction of data sets with specific realistic properties that could be used to demonstrate situations in which neural network analysis had particular strengths in comparison with Cox regression modeling. Using these simulated data sets neural network analysis could produce successful predictive models, find interactions between variables, and recognize the importance of variables that contributed to the hazard rate as a complex function of the variables value and in situations where the proportionality of hazards assumption was violated. It was also demonstrated that neural network analysis was not a 'black box ', but could lead to useful insights into the roles played by different prognostic variables in determining patient outcome.

AB - The purpose of this study was to demonstrate how a form of neural network analysis could be used to perform survival analysis on censored data, and to compare neural network analysis with the most commonly used technique for this type of analysis, Cox regression. In this study computer simulated data sets were used. The underlying rules connecting prognostic information to the hazard of death were defined to allow the construction of data sets with specific realistic properties that could be used to demonstrate situations in which neural network analysis had particular strengths in comparison with Cox regression modeling. Using these simulated data sets neural network analysis could produce successful predictive models, find interactions between variables, and recognize the importance of variables that contributed to the hazard rate as a complex function of the variables value and in situations where the proportionality of hazards assumption was violated. It was also demonstrated that neural network analysis was not a 'black box ', but could lead to useful insights into the roles played by different prognostic variables in determining patient outcome.

KW - Artificial intelligence

KW - Censored data

KW - Cox regression

KW - Neural networks

KW - Prognostic factors

UR - http://www.scopus.com/inward/record.url?scp=0028328448&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0028328448&partnerID=8YFLogxK

U2 - 10.1016/0304-3835(94)90095-7

DO - 10.1016/0304-3835(94)90095-7

M3 - Article

C2 - 8168059

AN - SCOPUS:0028328448

VL - 77

SP - 127

EP - 138

JO - Cancer Letters

JF - Cancer Letters

SN - 0304-3835

IS - 2-3

ER -