Enhanced Sensitivity BMI Calculator with Instant Notifications Utilizing XGBoost Models and SHAP Analysis

DOI

10.2991/978-94-6463-718-2_127How to use a DOI?

Keywords

XGBoost; SHAP; BMI prediction; machine learning; healthcare; interpretability; real-time prediction; personalized health; model generalization; overfitting prevention; health data

Abstract

This combination of XGboost and SHAP for BMI prediction provides tailored model implementation for different dataset augmentations, leading to high prediction accuracy and individual health recommendations. SHAP values also provide transparency and interpretability, allowing users to understand which factors most influence BMI predictions and increasing user trust in the system. The simplicity of complex models enabled with SHAP means access for non-expert users, which can be a game changer for uptake in healthcare. Computational Complexity: Although SHAP can be computationally heavy, this complexity inspires innovations that solve real-time BMI predictions with very low lag time. Operating on such vast amounts of data ensures greater generalizability, resulting in robust and reliable predictions over heterogeneous populations. The ability of XGBoost to prevent overfitting ensures that the model will generalize to new, unseen data, improving model performance. T h e models are also complex which have better capability to handle complex health data and thus provide more accurate, personalized BMI prediction. The BMI calculator remains up to date and in line with new data health trends thanks to regular updates to the model. Interpretation tools such as SHAP can help clinicians make data-driven decisions, which play a crucial role in the clinical context of prediction. Addressing scaling challenges drives the engineering of more computationally efficient solutions which allows the BMI calculator to work in low resource settings as well. Bias detection, analysis, and correction underpinned by sensitivity mean the model is more reliable and less biased. The model’s capacity to handle high-dimensional data encompasses a wide range of health indicators, and consequently, allows accurate BMI predictions across various users. Integrating multiple datasets from different healthcare systems increases scalability and enhances the calculator’s usability across distinct clinical settings. This ensures that the BMI calculator stays relevant with time, as new health trends emerge, and user data becomes more comprehensive. As the BMI calculator can incorporate multiple sources of data, it can be useful to a broad demographic, providing a versatile, general-purpose health tool. Demand for more specific health predictions has resulted in the model accounting for a greater number of health conditions and ensuring versatility for a variety of use cases. Designed for computational efficiency, we can further fine-tune the BMI calculator to run seamlessly on mobile or web interfaces, offering near real-time outputs that are especially critical for low-resource settings. This allows for the minimization of overfitting to individual conditions, which means the model is applicable across multiple health conditions. The models powered by deep learning algorithms increase the quality of predictions, which guarantees clients do a BMI estimation on precise time. Headline 10 — Providing the tool to learn, Drivers of Importance through SHAP Analysis: Driver of importance is a great insight offered through various algorithms — we discussed how SHAP analysis provides great detail on the relative importance of features to a prediction.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - M. Azhagesan
AU  - P. Palanisamy
AU  - D. Sathiya
AU  - R. G. Akshaiya
AU  - K. H. Gopika
AU  - S. Keerthana
PY  - 2025
DA  - 2025/05/23
TI  - Enhanced Sensitivity BMI Calculator with Instant Notifications Utilizing XGBoost Models and SHAP Analysis
BT  - Proceedings of the International Conference on Sustainability Innovation in Computing and Engineering (ICSICE 2024)
PB  - Atlantis Press
SP  - 1524
EP  - 1538
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6463-718-2_127
DO  - 10.2991/978-94-6463-718-2_127
ID  - Azhagesan2025
ER  -