A parallel concept learning algorithm based upon version space strategy

An attempt is made to apply the technique of parallel processing to concept learning. A parallel model suited for parallel learning and a parallel learning algorithm obtained from this model by applying the strategy of version space are proposed. The time complexity of the algorithm is compared with that of the sequential learning algorithm, and a satisfactory result is obtained-with N/2 processors, most learning problems can be processed in O(log N) time. The authors also study the influence of the arrangement of the training instances on the performance of their parallel algorithm. The final result shows that when the number of training instances is large, learning in parallel is worthwhile