Stream Data Classification with Hoeffding Tree : An Ensemble Learning Approach

Scientists around the world study data mining extensively, but many methods are limited to analyzing small databases. Technological advances have led to the emergence of Incremental Machine Learning and Stream Data Classification to handle large amounts of diverse data. The challenge is to quickly extract information from incoming sequences of data, but the high speed and complexity of the input data limit the application of previously proposed methods. The Hoeffding tree algorithm is crucial for Stream Data Classification and employs the Hoeffding bound to select a splitting feature. In this paper, we propose a method that combines an Incremental Decision Tree called the Hoeffding tree with Ensemble machine learning using bagging to enhance accuracy. Our implementation and analysis show that our proposed method improves accuracy compared to the simple Hoeffding tree. We also analyze the algorithm with different numbers of base models and examine graph diagrams to illustrate the improvement in accuracy.