An Investigative Study into Predictive Modeling for Early Detection of Eye Disorders: Challenges, Strategies and Mitigations

Authors

  • Ms. Vaishali Kale Research Scholar, D.Y Patil Institute of Management of Computer Applications and Management. Pune, Maharashtra, India. Author
  • Dr. Kavita Surawanshi Research Scholar, D.Y Patil Institute of Management of Computer Applications and Management. Pune, Maharashtra, India. Author

DOI:

https://doi.org/10.47392/IRJAEM.2024.0353

Keywords:

RF, Predictive modelling, LR, KNN, Early detection

Abstract

Eye disorders pose significant health challenges globally, affecting millions of individuals. Early detection is crucial for timely intervention and improved outcomes [1]. Our research investigates predictive modeling techniques to identify eye diseases at an early stage. We explore the impact of unique patient identifiers (Patient_ID), age (P_Age), and gender (P_Gender) on eye health. Understanding these factors is essential for personalized care and risk assessment. The study aims to bridge the gap between clinical practice and data-driven approaches, emphasizing the importance of accurate predictions for disease severity. To address our research objectives, we employ machine learning algorithms and statistical techniques. We leverage diagnostic information from both eyes using Left_Diagnostic_Keywords and Right_Diagnostic_Keywords. These keywords capture specific eye conditions, including glaucoma (G), cataracts (C), age-related macular degeneration (A), hypertensive retinopathy (H), macular edema (M), and optic neuritis (O). Our methodology integrates feature engineering, model selection, and cross-validation. We explore the trade-offs between sensitivity, specificity, and interpretability. Our findings reveal promising predictive capabilities for early detection. We examine machine learning algorithms to classify retinal images for glaucoma, diabetic retinopathy, and normal eyes, stressing the need for early detection. In our study, we compare the effectiveness of various machine learning methods. KNN proves to be the most effective for the given dataset. We focus on feature importance, analyzing attributes such as age, gender, diagnostic keywords, vision clarity, and specific conditions like glaucoma and cataracts to identify key predictors of eye diseases. By embracing data-driven insights,[2] we empower healthcare professionals to enhance patient outcomes and reduce the burden of eye diseases.

Downloads

Download data is not yet available.

Downloads

Published

2024-07-27