MLP-Based Lung Cancer Prediction and Feature Importance Evaluation

Zonglin Jiang

2024

Abstract

Lung cancer remains one of the deadliest cancers globally, with high mortality rates due to the challenges of early detection. Traditional diagnostic methods, such as CT scans and biopsies, have limitations, including the risk of human error and patient discomfort. With the advent of machine learning (ML) technologies, early detection has improved significantly. This paper investigates the importance of features in lung cancer prediction using a Random Forest model and a Multilayer Perceptron (MLP) model. The dataset used consists of 309 clinical samples and 15 features, with binary classification into cancerous and non-cancerous cases. After data preprocessing, the models were trained and evaluated to assess the contribution of different features. Age, Allergy, and Swallowing Difficulty were found to be the most important features in both models. The study highlights the impact of dataset imbalance on feature importance and model performance. Future work will focus on addressing this imbalance to improve prediction accuracy and reliability in clinical applications.

Paper Citation

in Harvard Style

Jiang Z. (2024). MLP-Based Lung Cancer Prediction and Feature Importance Evaluation. In Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM; ISBN 978-989-758-738-2, SciTePress, pages 320-323. DOI: 10.5220/0013329900004558

in Bibtex Style

@conference{mlscm24,
author={Zonglin Jiang},
title={MLP-Based Lung Cancer Prediction and Feature Importance Evaluation},
booktitle={Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM},
year={2024},
pages={320-323},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013329900004558},
isbn={978-989-758-738-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM
TI - MLP-Based Lung Cancer Prediction and Feature Importance Evaluation
SN - 978-989-758-738-2
AU - Jiang Z.
PY - 2024
SP - 320
EP - 323
DO - 10.5220/0013329900004558
PB - SciTePress