Early Prediction of Pancreatic Cancer Using Data Analysis Techniques in IBM SPSS Modeler

Principal Investigator

Name
Iñigo Monedero Goicoechea

Degrees
Ph.D.

Institution
University of Seville

Position Title
Tenure-track professor at University of Seville

Email
imonedero@us.es

About this CDAS Project

Study
PLCO (Learn more about this study)

Project ID
PLCO-1951

Initial CDAS Request Approval
Jul 11, 2025

Title
Early Prediction of Pancreatic Cancer Using Data Analysis Techniques in IBM SPSS Modeler

Summary
Pancreatic cancer is one of the most lethal oncological diseases due to its late diagnosis and the scarcity of early symptoms. This work seeks to apply data analysis techniques with IBM SPSS Modeler on clinical and patient demographic data to build a predictive system that identifies individuals at high risk of developing pancreatic cancer. The model can help prioritize more invasive and costly clinical tests only in patients with high probability of risk.

Aims

1. Acquire and analyze a relevant set of data, including factors associated with pancreatic cancer
2. Preprocess and clean the data with IBM SPSS Modeler (missing value treatment, discretization, normalization).
3. Explore relationships between variables to discover relevant patterns using graphical and statistical tools.
4. Build predictive models with algorithms such as: decision trees (C&R Tree, CHAID), Neural networks, Logistic regression and
SVM
5. Evaluate the performance of each model using cross-validation techniques.
6. Extract clinical rules and interpretations useful for medical decision making.
7. To propose a clinical decision support tool based on the trained model.

Collaborators

Iñigo Monedero Goicoechea: tenure-track professor at University of Seville