Machine Learning Algorithms for Predicting the Risk of Pancreatic Cancer

EasyChair Preprint 13587

Abstract

Pancreatic cancer is a devastating disease with a high mortality rate, emphasizing the need for early detection and accurate risk prediction. Machine learning algorithms have emerged as promising tools for predicting the risk of pancreatic cancer, leveraging clinical and demographic data to provide valuable insights. This abstract provides an overview of the key aspects involved in utilizing machine learning algorithms for pancreatic cancer risk prediction.

The process begins with data collection and preprocessing, involving the identification of relevant datasets and the application of cleaning techniques. Feature selection and extraction methods are employed to identify informative variables. Supervised learning algorithms, such as logistic regression, decision trees, random forests, support vector machines (SVM), and gradient boosting algorithms, are then utilized to build predictive models. These algorithms enable binary classification and offer interpretability, adaptability, and robustness.

Additionally, unsupervised learning algorithms, including clustering algorithms and dimensionality reduction techniques, are applied to identify subgroups and risk profiles within the dataset. This aids in further understanding the heterogeneity of pancreatic cancer and provides valuable insights into its risk factors.

Keyphrases: Confusion Matrix, Train-Test Split, cross-validation, evaluation metrics, model evaluation, model validation

@booklet{EasyChair:13587,
  author    = {Elizabeth Henry and Harold Jonathan},
  title     = {Machine Learning Algorithms for Predicting the Risk of Pancreatic Cancer},
  howpublished = {EasyChair Preprint 13587},
  year      = {EasyChair, 2024}}