Exploiting Discerning Classification Algorithms to Anticipate the Proneness to Diabetes in Its Nascent Phase

Early-stage diabetes risk prediction plays a crucial role in proactive healthcare management and prevention strategies. Prompt intervention and lifestyle change are made possible by identifying individuals with a heightened risk of developing diabetes, thereby improving patient outcomes. This research paper focuses on predicting diabetes by identifying important attributes and exploring the relationships between these attributes. For the objective of predicting diabetes, various classification algorithms are implemented using the Weka tool. In this study, we analyze the use of classification techniques for early-stage diabetes risk prediction. To begin with, we choose seventeen features viz., the attributes include Age, Gender. Frequent urination, Increased thirst, unexpected reduction in weight, Fatigue, Excessive hunger, Fungal infection, Blurred vision, Skin itching, restlessness, Slow healing of wounds, Partial weakness, Stiff muscles, Hair loss, Overweight, Class. Moreover, machine learning classification algorithms, notably J48, Random Forest along with REPtree are employed to identify early-stage diabetes. We evaluate the performance of various classification algorithms to precisely forecast the risk of diabetes onset in individuals. In conclusion, the Random Forest classifier approach outperforms the other two approaches.