摘要
针对传统的乒乓球视频数据挖掘算法存在算法效率不高、优化分类性能差的问题,提出了一种基于改进人工鱼群算法的乒乓球视频数据挖掘模型。首先,对人工鱼群算法变量的取值范围进行优化。然后在原有人工鱼行为的基础上增加最优的人工鱼状态信息,从而引导人工鱼快速接近全局最优,提高算法的速度。最后,对传统人工鱼群算法的搜索策略进行自适应优化,并将改进算法应用于乒乓球比赛视频数据挖掘。仿真实验表明,基于改进人工鱼群算法的乒乓球比赛视频数据挖掘模型具有较好的收敛效率和分类能力,可以在乒乓球比赛视频中挖掘出更多的图像属性和击球类型。
Aiming at the problems of poor performance and poor classification ability of traditional table tennis video data mining algorithms, a new model of table tennis video data mining based on improved artificial fish school algorithm is proposed. First, this paper optimizes the range of the algorithm variables of artificial fish school. Then, based on the original artificial fish behavior, the optimal artificial fish status information is added to guide the artificial fish to quickly approach the global optimum and improve the speed of the algorithm. Finally, the search strategy of traditional artificial fish swarm algorithm is adaptively optimized, and the improved algorithm is applied to the video data mining of table tennis competition. The simulation results show that the video data mining model based on improved artificial fish swarm algorithm has better convergence efficiency and classification ability, and more image attributes and shot types can be excavated in the table tennis game video.
作者
任祥钰
REN Xiang-yu(Xi'an Medical University, Xi'an 710021 China)
出处
《自动化技术与应用》
2019年第5期17-21,共5页
Techniques of Automation and Applications
关键词
改进的人工鱼群算法
乒乓球比赛
数据挖掘
行为优化
自适应策略
improved artificial fish swarm algorithm
table tennis competition
data mining
behavior optimization
adaptive strategy
