A total of 126 traditional villages in Guangdong were selected for research to analyze their spatial distribution pattern, spatial autocorrelation, and Han subgroup characteristics through the spatial analysis with GI...A total of 126 traditional villages in Guangdong were selected for research to analyze their spatial distribution pattern, spatial autocorrelation, and Han subgroup characteristics through the spatial analysis with GIS. The results showed that the traditional villages in Guangdong were distributed compactly, which was similar to that in other provinces; on the prefecture-level city scale, they were mainly distributed in Meizhou, Qingyuan, Zhanjiang, and Zhaoqing, showing a central distribution; on the regional scale, they were mainly distributed in northern Guangdong and the Pearl River Delta, indicating an unbalanced distribution; their density of spatial distribution showed a significant difference with a provincial distribution density of 0.59 per thousand km^2 and the highest distribution density of 2.14 per thousand km^2; there were four gathering areas accommodating many traditional villages, namely Meizhou-Chaoshan area, Guangfo-Dongguan area, Leizhou in Zhanjiang, and Lianzhou in Qingyuan; the spatial distribution of traditional villages, which presented a pattern of hot spot—sub-hot spot—sub-cold spot—cold spot, had a significant global spatial autocorrelation, and areas in similar sizes were tending towards gathering; traditional villages in Guangdong belonged to the Cantonese-speaking group and Hakka-speaking group with a density of 0.59 per thousand km^2 and 0.48 per thousand km^2 respectively, and Han subgroup with the highest distribution density was the minority areas of 1.31 per thousand km^2, followed by Shaozhou dialect area with a density of 1.00 per thousand km^2.展开更多
Most of the neural network handwritten digits recognition systems adopt a network with single structure. In this paper,we implement a subgroup neural network structure based on feature description to recognize handwr...Most of the neural network handwritten digits recognition systems adopt a network with single structure. In this paper,we implement a subgroup neural network structure based on feature description to recognize handwritten digits. It is noticeable that this subgroup network is a hybrid system. The experiment result indicates that the subgroup network gets much better performance than usual single network Structure.展开更多
基金Sponsored by Guangdong Provincial Natural Science Foundation(2015A030313852)Guangdong Provincial Scientific and Technological Programs(2014A020219006,2015B070701020)
文摘A total of 126 traditional villages in Guangdong were selected for research to analyze their spatial distribution pattern, spatial autocorrelation, and Han subgroup characteristics through the spatial analysis with GIS. The results showed that the traditional villages in Guangdong were distributed compactly, which was similar to that in other provinces; on the prefecture-level city scale, they were mainly distributed in Meizhou, Qingyuan, Zhanjiang, and Zhaoqing, showing a central distribution; on the regional scale, they were mainly distributed in northern Guangdong and the Pearl River Delta, indicating an unbalanced distribution; their density of spatial distribution showed a significant difference with a provincial distribution density of 0.59 per thousand km^2 and the highest distribution density of 2.14 per thousand km^2; there were four gathering areas accommodating many traditional villages, namely Meizhou-Chaoshan area, Guangfo-Dongguan area, Leizhou in Zhanjiang, and Lianzhou in Qingyuan; the spatial distribution of traditional villages, which presented a pattern of hot spot—sub-hot spot—sub-cold spot—cold spot, had a significant global spatial autocorrelation, and areas in similar sizes were tending towards gathering; traditional villages in Guangdong belonged to the Cantonese-speaking group and Hakka-speaking group with a density of 0.59 per thousand km^2 and 0.48 per thousand km^2 respectively, and Han subgroup with the highest distribution density was the minority areas of 1.31 per thousand km^2, followed by Shaozhou dialect area with a density of 1.00 per thousand km^2.
文摘Most of the neural network handwritten digits recognition systems adopt a network with single structure. In this paper,we implement a subgroup neural network structure based on feature description to recognize handwritten digits. It is noticeable that this subgroup network is a hybrid system. The experiment result indicates that the subgroup network gets much better performance than usual single network Structure.
