适用于新媒体事件聚类模型的混合算法研究

摘要

To distinguish between new media event timely and accurately, based on analyzing the clustering algorithm usually used, the genetic algorithm and K_ Means are combined effectively. A clustering fitting new media event that called genetic algorithm with K_ Means is proposed The slow convergence of genetic algorithm is improved, and the problem of easy to fall into local optimal solution of K _ Means is improved at the same time The operations of optimal preservation strategies, single-point crossover and single-point mutation are used; it can ensure the convergence of genetic algorithm with K _ Means. Simulation analysis demonstrated the feasibility and the effectiveness of the algorithm. A new research method is provided for the clustering of new media events.%为了及时准确区分新媒体事件,在分析常用聚类算法的基础上,将遗传算法与K_Means算法有效结合,提出了一种适合运用于新媒体事件聚类的混合K均值遗传算法.该算法同时改善了遗传算法的收敛较慢,以及K_ Means算法易于陷入局部最优解的不足,并运用最优保存策略、单点交叉以及单点变异操作,更大程度保证了混合K均值遗传算法的收敛.仿真实验表明了算法的可行性和有效性,为新媒体事件的聚类研究提供了新的研究方法.