Diabetes Clinical Dataset from UITH and General Hospital Offa, Nigeria

Description

Data is anonymized clinical records for the purposes of diabetes diagnosis research from two tertiary health care centres in Nigeria. The total number of patients used was 800 from the University of Ilorin Teaching Hospital and General Hospital Offa. After data cleaning and data pre-processing, 768 full data sets were kept. Clinical parameters that are routinely used in the diagnosis of diabetes are included in the data set, such as number of pregnancies, plasma glucose concentration, diastolic blood pressure, skin thickness, serum insulin level, body mass index (BMI), diabetes pedigree function, and age. The outcome variable is the diabetic status (diabetic = 1, non-diabetic = 0). The data is created to facilitate machine learning studies in the diagnosis of diabetes and for the specific clinical use of localized African clinical data for predictive modeling studies. Patient data has been anonymised and no personal data has been given.

Steps to reproduce

1. Load the CSV dataset into Python, R, or any statistical software. 2. Perform data inspection to identify missing or inconsistent values. 3. Remove incomplete records and invalid entries. 4. Normalize or standardize numerical variables before model training. 5. Split the dataset into training and testing sets using ratios such as 70:30, 75:25, or 80:20. 6. Train machine learning models including Decision Tree, Support Vector Machine, and Artificial Neural Network. 7. Evaluate model performance using accuracy, sensitivity, specificity, and precision metrics. The target variable represents diabetes status, where 1 indicates diabetic patients and 0 indicates non-diabetic patients.

Institutions

• Osun State University

  Osun State, Osogbo

Categories

Licence