A Fast Calculation of Metric Scores for Learning Bayesian Network

Frequent counting is a very so often required operation in machine learning algorithms. A typical machine learning task, learning structure of Bayesian network (BN for short) based on metric scoring, is introduced as an example task which heavily relies on frequent counting. A fast calculation method for frequent counting enhanced with two cache layers is then presented for learning BN . We name this speedup technique RC solution. The main contribution of RC is to eliminate comparison operations for frequent counting by introducing a multi-radix number system calculation. Both mathematical analysis and empirical comparison between RC and state-of-the-art solution are conducted. The results show that RC is dominant prior over the current state-of-the-art solution at least in solving the problem of learning BN . Further discussions on how to extend RC to other similar learning tasks are also developed at the end of this paper.