Since most parameter control methods are based on prior knowledge, it is difficult for them to solve various problems.In this paper, an adaptive selection method used for operators and parameters is proposed and named...Since most parameter control methods are based on prior knowledge, it is difficult for them to solve various problems.In this paper, an adaptive selection method used for operators and parameters is proposed and named double adaptive selection(DAS) strategy. Firstly, some experiments about the operator search ability are given and the performance of operators with different donate vectors is analyzed. Then, DAS is presented by inducing the upper confidence bound strategy, which chooses suitable combination of operators and donates sets to optimize solutions without prior knowledge. Finally, the DAS is used under the framework of the multi-objective evolutionary algorithm based on decomposition, and the multi-objective evolutionary algorithm based on DAS(MOEA/D-DAS) is compared to state-of-the-art MOEAs. Simulation results validate that the MOEA/D-DAS could select the suitable combination of operators and donate sets to optimize problems and the proposed algorithm has better convergence and distribution.展开更多
One of the challenges the world is facing today is global climate change and its associated impacts. Uganda is a climate sensitive country with over 90% of the population dependent on climate sensitive sectors.Gradual...One of the challenges the world is facing today is global climate change and its associated impacts. Uganda is a climate sensitive country with over 90% of the population dependent on climate sensitive sectors.Gradual and sudden variations in climatic parameters will therefore render the livelihoods of Ugandans very vulnerable.A number of approaches to enhance the resilience of communities to展开更多
To realize content-hased retrieval of large image databases, it is required to develop an efficient index and retrieval scheme. This paper proposes an index algorithm of clustering called CMA, which supports fast retr...To realize content-hased retrieval of large image databases, it is required to develop an efficient index and retrieval scheme. This paper proposes an index algorithm of clustering called CMA, which supports fast retrieval of large image databases. CMA takes advantages of k-means and self-adaptive algorithms. It is simple and works without any user interactions. There are two main stages in this algorithm. In the first stage, it classifies images in a database into several clusters, and automatically gets the necessary parameters for the next stage-k-means iteration. The CMA algorithm is tested on a large database of more than ten thousand images and compare it with k-means algorithm. Experimental results show that this algorithm is effective in both precision and retrieval time.展开更多
This paper presents a new Section Set Adaptive FCM algorithm.The algorithm solved the shortcomings of local optimality,unsure classification and clustering numbers ascertained previously.And it improved on the archite...This paper presents a new Section Set Adaptive FCM algorithm.The algorithm solved the shortcomings of local optimality,unsure classification and clustering numbers ascertained previously.And it improved on the architecture of FCM al- gorithm,enhanced the analysis for effective clustering.During the clustering processing,it may adjust clustering numbers dy- namically.Finally,it used the method of section set decreasing the time of classification.By experiments,the algorithm can im- prove dependability of clustering and correctness of classification.展开更多
针对利用海量数据构建分类模型时训练数据规模大、训练时间长且碳排放量大的问题,提出面向低能耗高性能的分类器两阶段数据选择方法TSDS(Two-Stage Data Selection)。首先,通过修正余弦相似度确定聚类中心,并将样本数据进行基于不相似...针对利用海量数据构建分类模型时训练数据规模大、训练时间长且碳排放量大的问题,提出面向低能耗高性能的分类器两阶段数据选择方法TSDS(Two-Stage Data Selection)。首先,通过修正余弦相似度确定聚类中心,并将样本数据进行基于不相似点的分裂层次聚类;其次,对聚类结果按数据分布自适应抽样以组成高质量的子样本集;最后,利用子样本集在分类模型上训练,在加速训练过程的同时提升模型精度。在Spambase、Bupa和Phoneme等6个数据集上构建支持向量机(SVM)和多层感知机(MLP)分类模型,验证TSDS的性能。实验结果表明在样本数据压缩比达到85.00%的情况下,TSDS能将分类模型准确率提升3~10个百分点,同时加速模型训练,使训练SVM分类器的能耗平均降低93.76%,训练MLP分类器的能耗平均降低75.41%。可见,TSDS在大数据场景的分类任务上既能缩短训练时间和减少能耗,又能提升分类器性能,从而助力实现“双碳”目标。展开更多
基金supported by the National Natural Science Foundation of China(7177121671701209)
文摘Since most parameter control methods are based on prior knowledge, it is difficult for them to solve various problems.In this paper, an adaptive selection method used for operators and parameters is proposed and named double adaptive selection(DAS) strategy. Firstly, some experiments about the operator search ability are given and the performance of operators with different donate vectors is analyzed. Then, DAS is presented by inducing the upper confidence bound strategy, which chooses suitable combination of operators and donates sets to optimize solutions without prior knowledge. Finally, the DAS is used under the framework of the multi-objective evolutionary algorithm based on decomposition, and the multi-objective evolutionary algorithm based on DAS(MOEA/D-DAS) is compared to state-of-the-art MOEAs. Simulation results validate that the MOEA/D-DAS could select the suitable combination of operators and donate sets to optimize problems and the proposed algorithm has better convergence and distribution.
文摘One of the challenges the world is facing today is global climate change and its associated impacts. Uganda is a climate sensitive country with over 90% of the population dependent on climate sensitive sectors.Gradual and sudden variations in climatic parameters will therefore render the livelihoods of Ugandans very vulnerable.A number of approaches to enhance the resilience of communities to
基金This project was supported by National High Tech Foundation of 863 (2001AA115123)
文摘To realize content-hased retrieval of large image databases, it is required to develop an efficient index and retrieval scheme. This paper proposes an index algorithm of clustering called CMA, which supports fast retrieval of large image databases. CMA takes advantages of k-means and self-adaptive algorithms. It is simple and works without any user interactions. There are two main stages in this algorithm. In the first stage, it classifies images in a database into several clusters, and automatically gets the necessary parameters for the next stage-k-means iteration. The CMA algorithm is tested on a large database of more than ten thousand images and compare it with k-means algorithm. Experimental results show that this algorithm is effective in both precision and retrieval time.
基金Science and Researching Foundation of Jiamusi University(L2006-12)
文摘This paper presents a new Section Set Adaptive FCM algorithm.The algorithm solved the shortcomings of local optimality,unsure classification and clustering numbers ascertained previously.And it improved on the architecture of FCM al- gorithm,enhanced the analysis for effective clustering.During the clustering processing,it may adjust clustering numbers dy- namically.Finally,it used the method of section set decreasing the time of classification.By experiments,the algorithm can im- prove dependability of clustering and correctness of classification.
文摘在高压并联电抗器声纹信号监测系统中,长时海量无标签声纹的高维非平稳性导致特征提取困难、无监督聚类适应性差。由此提出了一种基于深度自适应K-means++算法(deep adaptive K-means++clustering algorithm,DAKCA)的750 kV电抗器声纹聚类方法。首先通过采用两阶段无监督策略微调的改进堆叠稀疏自编码器(stacked sparse autoencoder,SSAE),对快速傅里叶变换后的归一化频域数据提取电抗器原始声纹32维深度特征。进一步提出了依据最近邻聚类有效性指标(clustering validation index based on nearest neighbors,CVNN)的自适应K-means++聚类算法,构建了能自适应确定最优聚类个数的电抗器声纹聚类模型。最后通过西北地区某750 kV电抗器实测声纹数据集进行了验证。结果表明,DAKCA算法对无标签声纹数据在不同样本均衡程度下能够稳定提取32维深度特征,并实现最优聚类,为直接高效利用电抗器无标签声纹数据提供了参考。
文摘针对利用海量数据构建分类模型时训练数据规模大、训练时间长且碳排放量大的问题,提出面向低能耗高性能的分类器两阶段数据选择方法TSDS(Two-Stage Data Selection)。首先,通过修正余弦相似度确定聚类中心,并将样本数据进行基于不相似点的分裂层次聚类;其次,对聚类结果按数据分布自适应抽样以组成高质量的子样本集;最后,利用子样本集在分类模型上训练,在加速训练过程的同时提升模型精度。在Spambase、Bupa和Phoneme等6个数据集上构建支持向量机(SVM)和多层感知机(MLP)分类模型,验证TSDS的性能。实验结果表明在样本数据压缩比达到85.00%的情况下,TSDS能将分类模型准确率提升3~10个百分点,同时加速模型训练,使训练SVM分类器的能耗平均降低93.76%,训练MLP分类器的能耗平均降低75.41%。可见,TSDS在大数据场景的分类任务上既能缩短训练时间和减少能耗,又能提升分类器性能,从而助力实现“双碳”目标。
