...问题并不一定是故意欺骗，因为在数据采集的实践中确实存在样本量失衡的情况，如果遇到这种情况一般使用欠抽样（Under-sampling）和过抽样（Oversampling）进行样本平衡。通常来说样本的问题主要分为以下三种情况：
基于6个网页-相关网页
The thesis describes a solution for software short-wave receiver. Analog is converted to digital signal by using IF under-sampling technology.
本文阐述了一种软件无线电短波接收机的实现方案，它的模拟域到数字域的转换是采用欠采样技术和中频数字化实现的。
参考来源 - 短波单边带软件无线电DSP模块的研制Frequency domain detect mainly through fast Fourier transform to get the carrier frequency and modulation frequency. In order to meet the resolution we adopt the Under-Sampling and ZFFT algorithm.
频谱分析主要通过对移频信号进行快速傅里叶变换求得主峰处频率(载频频率)及主峰与次峰的差值(调制频率),其中为了满足分辨率的要求采用了欠采样和ZFFT算法。
参考来源 - ZPW·2,447,543篇论文数据，部分数据来源于NoteExpress
以上来源于: WordNet
This paper discussed the uniform under-sampling of multiple bandpass signals in digital receiver.
该文讨论了宽带数字接收机中对多个复或实的带通信号的直接均匀采样。
Sampling methods including over-sampling, under-sampling were taken to make up the shortage caused by the imbalanced data.
采用过抽样、欠抽样技术来弥补数据类别分布不平衡对模型性能的不利影响。
To test your null hypothesis, you want to know the probability of obtaining a value this extreme under the assumption that it is due to random sampling variability.
要检验虚假设，需要知道在假设存在随机抽样可变性的情况下获得这么一个极限值的概率。
应用推荐