The track association problem of radar and electronic support measure （ESM） has been considered in the literature for several years. This problem is crucial for radar-to-ESM track fusion and is complicated by the pr...The track association problem of radar and electronic support measure （ESM） has been considered in the literature for several years. This problem is crucial for radar-to-ESM track fusion and is complicated by the presence of individual systematic errors and measurement errors. In order to improve the track association of radar and ESM sensors, a pseudo-linear filtering algorithm is proposed to estimate the target states and improve the stability of the filter. It is found that, however, the correct probability of radar- to-ESM track association decreases as the radar measurement error decreases, when the pseudo-linear filter is used for ESM sensor filtering. In view of the strange phenomenon, this paper analyzes the reason for it by using the statistic theory and further performs Monte Carlo simulation to verify the analysis.展开更多
随着互联网技术的高速发展,人们与数字音乐的关系更加紧密,人们会依据自己的偏好以及场所选择音乐,因此如何有效管理数量庞大的音乐并对其分门别类显得尤为重要。为提高音乐分类的准确率,本文从音频中提取特征向量,运用遗传算法优化支...随着互联网技术的高速发展,人们与数字音乐的关系更加紧密,人们会依据自己的偏好以及场所选择音乐,因此如何有效管理数量庞大的音乐并对其分门别类显得尤为重要。为提高音乐分类的准确率,本文从音频中提取特征向量,运用遗传算法优化支持向量机实现音乐流派分类;从歌词中提取特征关键词,采用L A S SO降维实现文本情感分类,最终构建双模态音乐分类模型。结果表明,该分类方法准确率为73.1%,可靠性与稳定性良好,有效地避免了传统方法产生局部最优的问题。展开更多
基金supported by the National Natural Science Foundation of China (609721596103200161002006)
文摘The track association problem of radar and electronic support measure （ESM） has been considered in the literature for several years. This problem is crucial for radar-to-ESM track fusion and is complicated by the presence of individual systematic errors and measurement errors. In order to improve the track association of radar and ESM sensors, a pseudo-linear filtering algorithm is proposed to estimate the target states and improve the stability of the filter. It is found that, however, the correct probability of radar- to-ESM track association decreases as the radar measurement error decreases, when the pseudo-linear filter is used for ESM sensor filtering. In view of the strange phenomenon, this paper analyzes the reason for it by using the statistic theory and further performs Monte Carlo simulation to verify the analysis.
文摘随着互联网技术的高速发展,人们与数字音乐的关系更加紧密,人们会依据自己的偏好以及场所选择音乐,因此如何有效管理数量庞大的音乐并对其分门别类显得尤为重要。为提高音乐分类的准确率,本文从音频中提取特征向量,运用遗传算法优化支持向量机实现音乐流派分类;从歌词中提取特征关键词,采用L A S SO降维实现文本情感分类,最终构建双模态音乐分类模型。结果表明,该分类方法准确率为73.1%,可靠性与稳定性良好,有效地避免了传统方法产生局部最优的问题。
