Objective <p>Diabetes is one of the most common chronic health threats worldwide. Early detection of diabetes is difficult and diagnosis and treatment processes can be costly. Data mining techniques offer powerful tools for predictive analysis and knowledge extraction from large data sets. This study aims to identify symptoms that collectively influence the development of diabetes by data mining and identify risk parameters for early detection.</p> Materials and methods <p>The study uses a dataset of 520 patient records collected from Sylhet Diabetes Hospital in Sylhet, Bangladesh. This dataset is based on real-world data from the UCI Machine Learning Repository. The Apriori algorithm, which is widely used in data mining, was applied to analyze the symptoms associated with diabetes using association analysis. The algorithm analyzed the relationships between symptoms based on support, confidence and lift values.</p> Results <p>The analysis identified eight key symptoms that significantly contribute to diabetes risk when they occur together: gender, polyuria, polydipsia, sudden weight loss, weakness, blurred vision, partial paresis and obesity. The co-occurrence of these symptoms increases the likelihood of developing diabetes by 1.63 times. These findings emphasize the importance of assessing symptoms collectively rather than in isolation.</p> Conclusion <p>The results of the study emphasize the importance of individuals at risk of diabetes and healthcare professionals to monitor these symptoms and take necessary precautions. The study shows that association rule mining, especially the Apriori algorithm, is a valuable tool for identifying symptom associations and facilitating early diabetes detection. The findings will contribute to early detection of diabetes and prevention of complications related to the disease through simple symptom analysis.</p>