Chronic Kidney Diseases Predications Using Machine Learning

Abstract

Kidney Diseases Are An Increasingly Important Global Health Problem For Millions Of People A Year, For Which Early Detection And Accurate Prediction Play A Major Role In Improving Patient Outcomes And Reducing The Burden On Healthcare Systems. This Project Seeks To Develop A Predictive Model Of Kidney Disease Diagnosis Based On Machine Learning That Assumes Control Over Clinical Data Through Highly Advanced Analytical Techniques. The Dataset Was Preprocessed So That Missing Values Were Handled, Features Standardized, And The Most Relevant Predictors Chosen. Exploratory Data Analysis (EDA) Was Also Conducted To Look For Any Type Of Patterns And Relationship Among Clinical Attributes. Two Machine Learning Algorithms Were Used For The Construction Of The Predictive Models - Random Forest And Decision Tree. Training And Validation Of Models Through A Structured Approach Ensure Robust Evaluation Through Metrics Such As Accuracy, Precision, Recall, And Confusion Matrices. Data Visualization Techniques Were Also Employed To Enhance Interpretability And Generate Actionable Insights In The Dataset. Results The Results Have Shown That The Implemented Models Work Well, With The Random Forest Algorithm Performing Best In Terms Of Accuracy And Reliability. Therefore, This Project Serves As A Possible Application Of ML In The Clinical Field To Alert Healthcare Professionals When Kidney Diseases Would Commence. This Early Detection Could Ultimately Serve To Improve A Patient's Situation While Providing An Interface For Predictive Analytics In Clinical Decision-support Systems. The Accuracy Of Light GBM Is 95%. XG Boost Is 94%.

Keywords