首页|Findings from Naval University of Engineering Yields New Findings on Machine Lea rning (Cross-validation Strategy for Performance Evaluation of Machine Learning Algorithms In Underwater Acoustic Target Recognition)
Findings from Naval University of Engineering Yields New Findings on Machine Lea rning (Cross-validation Strategy for Performance Evaluation of Machine Learning Algorithms In Underwater Acoustic Target Recognition)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News-Data detailed on Machine Learning have been prese nted. According to news reporting originating in Wuhan, People's Republic of Chi na, by NewsRx journalists, research stated, "The performance evaluation of learn ing algorithms is essential in underwater acoustic target recognition. Underwate r acoustic data typically show temporal structures." The news reporters obtained a quote from the research from the Naval University of Engineering, "However, these structures are frequently ignored when evaluatin g the performance of learning algorithms, resulting in overestimating predictive accuracy. The similarity analysis of underwater acoustic samples indicates clus ter structures in the feature space, which follow a manifold distribution over t ime. A uniform block cross-validation method and a clustering block crossvalidat ion method are proposed to evaluate the performance of learning algorithms. The effect of block size is investigated using the simulated and the real data. The results indicate that the proposed clustering block crossvalidation method is su itable for evaluating algorithms when interpolation is prediction objective only . The uniform block cross-validation is suitable for evaluating algorithms ‘ int erpolation and extrapolation abilities. Moreover, the proposed two methods are s uperior to the random cross-validation method."
WuhanPeople's Republic of ChinaAsiaAlgorithmsCyborgsEmerging TechnologiesMachine LearningNaval University of Engineering
NETL
NSTL
