Selected WorkNAFLD-THESIS-EARLY-LIVER-INJURY

NAFLD Thesis: Early Prediction of Liver Injury

Master’s thesis project: early prediction of liver injury signals in NAFLD using longitudinal clinical data + ML. Work in progress;

Hybrid MLExpert SystemsMLClinical DataTime SeriesFeature Engineering
Cover
Stack

Tech Stack

Pythonpandas/NumPyscikit-learnPyTorch (planned/used)Time-series feature engineeringExplainability (planned/used)
SectionOverview

Overview

A thesis-driven ML project focused on predicting early liver injury patterns in NAFLD using longitudinal patient data with clinically meaningful evaluation and interpretability.

SectionThe Problem

The Problem

NAFLD progression and liver injury risk emerge gradually and unevenly across time; early signals are subtle, data is messy, and outputs must be defensible for clinical use.

SectionThe Solution

The Solution

Building a data-to-model pipeline for longitudinal labs and patient context, engineering clinically grounded features, training baseline + sequence models, and validating with interpretable outputs suitable for a thesis setting.

SectionResults

Results & Impact

In progress , thesis work focused on robust data pipelines, model baselines, and clinically interpretable evaluation.