Handwritten Digit Recognition Using Ensemble Learning

In pattern recognition, the recognition of handwritten digits has always been a very challenging and tedious task. In this work, a simple novel approach is proposed to recognize the handwritten digits. The primary goal of this work is recognition of the handwritten digits by using ensemble learning. Ensemble learning improves convergence by decreasing the complexity of the model. The distribution of data in the random split and class-wise split for the base learners has been studied. Detailed analysis of how the load is distributed among the base learners and how it impacts the model accuracy and training time is also discussed. The overall trends of the ensemble model have also been analyzed.