论文部分内容阅读
主要研究了多噪声源共同作用下的混合噪声烦恼度的评价过程与预测方法.首先,设计并完成了固定播放时长噪声样本作用下的烦恼度主观评价实验,获得了人工合成的混合噪声样本作用下的混合噪声烦恼度(亦称总烦恼度)α_T评价数据与构成混合噪声样本的所有单一噪声样本单独作用时的烦恼度α_i(i=1,2,3,…,K;K为混合噪声样本中单一噪声样本的总数)评价数据.随后,细致分析了两组评价数据之间的关系,提出在已知α_i的基础上利用多元线性回归模型预测α_T.最后,解决了如何确定模型中对应各α_i的权值w_i(i=1,2,3,…,K)的问题.研究表明,以所提出的权值确定方法建立的多元线性回归预测模型能够较为成功地预测混合噪声样本作用下的总烦恼度评价值.
This paper mainly studies the evaluation process and prediction method of mixed noise annoyance under the combined action of multiple noise sources.Firstly, we designed and completed the annoyance subjective evaluation experiment under the effect of long-time noise samples of fixed broadcast, and obtained the synthetic mixed noise sample effect Α_T (i = 1, 2, 3, ..., K); where K is the noise of mixed noise (also referred to as the total annoyance) α_T is the degree of annoyance α_i (i = 1,2,3, ..., K) when all the single noise samples in the mixed noise sample are evaluated. The total number of single noise samples in the sample) evaluation data.Afterwards, the relationship between the two sets of evaluation data was carefully analyzed, and the prediction of α_T based on the known α_i was proposed.Finally, how to determine the corresponding The weights wi (i = 1, 2, 3, ..., K) of each α_i are studied.The research shows that the multivariate linear regression prediction model established by the proposed weighting method can predict the mixed noise samples more successfully The total annoyance rating.