An Algorithm of Alternately Mining Frequent Neighboring Class Set

Aiming to these present frequent neighboring class set mining algorithms existing more repeated computing and redundancy neighboring class set, this paper proposes an algorithm of alternately mining frequent neighboring class set, which is suitable for mining frequent neighboring class set of objects in large spatial data. The algorithm uses the regression method to create database of neighboring class set, and uses the alternative method to generate candidate frequent neighboring class set, namely, it uses increasing sequence to generate candidate in the one hand, it also uses decreasing sequence to generate candidate on the other hand, it only need scan once database to extract frequent neighboring class set. The algorithm improves mining efficiency by the alternative method, since not only using numerical variable to generate candidate is simple, but also using logic operation to compute support is very simple. The result of experiment indicates that the algorithm is faster and more efficient than presented algorithms when mining frequent neighboring class set in large spatial data.