A classifier ensemble is a combination of multiple diverse classifier models whose outputs are aggregated into a single prediction. Ensembles have been repeatedly shown to perform better than single classifier models, however, existing approaches trade off performance and robustness to class label noise. The objective of this paper is to first introduce a new perspective on multi-class ensemble classification by considering the training of the ensemble as a state estimation problem. The new perspective considers the final ensemble classifier model as a static state, which can be estimated using a Kalman filter that combines noisy estimates made by individual classifier models. A new algorithm based on this perspective, Kalman Filter-based Heuristic Ensemble (KFHE), is also presented in this paper which shows the practical applicability of the new perspective. Experiments performed on 30 real-life datasets, comparing KFHE with state-of-the-art multi-class ensemble classification algorithms uncover the potential and effectiveness of the proposed new perspective and algorithm. KFHE is shown to be significantly better or at least as good as the state-of-the-art algorithms for datasets both with and without class label noise.