期刊文献+
共找到2篇文章
< 1 >
每页显示 20 50 100
层次聚类和长短期记忆网络(LSTM)混合机器学习模型的数据资产估值模型
1
作者 彭守斌 杨学军 《计算机科学与应用》 2025年第6期45-55,共11页
研究数据资产估值在宏观层面能为数字经济发展、资源配置优化、行业规范和国家竞争力提升等方面提供支撑,对推动社会经济数字化转型意义重大。在数据资产估值过程中,会遇到诸多复杂问题,如数据异质性问题、时间序列特征挖掘问题、数据... 研究数据资产估值在宏观层面能为数字经济发展、资源配置优化、行业规范和国家竞争力提升等方面提供支撑,对推动社会经济数字化转型意义重大。在数据资产估值过程中,会遇到诸多复杂问题,如数据异质性问题、时间序列特征挖掘问题、数据维度高和复杂性问题、缺乏通用估值标准问题和突发外部事件影响问题。层次聚类和长短期记忆网络(LSTM)混合机器学习模型将层次聚类划分异质数据成簇,LSTM挖掘各簇时间序列特征,应对高维复杂数据,结合制定估值标准,快速适应突发变化;可有效应对以上诸多复杂问题。 展开更多
关键词 层次聚类 长短期记忆网络(LSTM) agglomerativeclustering (层次聚类算法) 门控机制(LSTM) 梯度消失/爆炸(RNN缺陷) 泛化能力 经典估值模型(成本法/市场法/收益法) 单一估值模型(K-Means/MLP/RNN)
在线阅读 下载PDF
A study on interaspecific biodiversity of eight groups of silkworm (Bombyx mori) by biochemical markers
2
作者 KAYVANETEBARI S.Z.MIRHOSEINI L.MATINDOOST 《Insect Science》 SCIE CAS CSCD 2005年第2期87-94,共8页
The recognition of biodiversity in different races and lines of silkworm (Bombyx mori) is very useful for breeding programs and production of high efficiency hybrids. In this study eight groups of silkworm were select... The recognition of biodiversity in different races and lines of silkworm (Bombyx mori) is very useful for breeding programs and production of high efficiency hybrids. In this study eight groups of silkworm were selected including 103, 107, Xihang 1 and 2 of Japanese origin and 104, 110, Koming 1 and 2 of Chinese origin. The activity levels of three enzymes including alkaline phosphatase, alanine aminotransferase and aspartate aminotransferase in haemolymph of fifth instar larva were measured. Moreover, the quantitative amount of total protein, cholesterol and glucose of haemolymph was evaluated.The data reveal that the activity level of measured macromolecules except for alkaline phosphatase were significantly different in all the groups. Hierarchical agglomerative clustering under UPGMA model separated line 104 from other groups. Two groups of Koming 1 and Xihang 1 had the most intergroup similarities. 展开更多
关键词 SILKWORM biochemical markers BIODIVERSITY hierarchical agglomerativeclustering
原文传递
上一页 1 下一页 到第
使用帮助 返回顶部