青海大学硕士学位论文
摘要
摘 要
自 2018 年以来，中美贸易摩擦持续升温。2018 年 7 月 10 日，美国开始挥
起贸易大棒，决定对来自中国的 2000 亿美元的商品加征关税，增加幅度达到
10%，并随之公布了相关商品清单，在这份关税清单中可以看到稀土资源被纳入
了其中，包含了稀土终端产品在内的稀土类氧化物、化合物等各种以稀土为原料
的相关商品。然而，在两个月后的 9月 17日美国持续加码，再次发布加税清单，
决定对来自中国的另外 2000亿美元商品加征 25%关税，第二年 5月 14日美国第
三次发布加税清单，决定对 3000 亿美元商品加征关税，但是在这两次贸易清单
中稀土却得到了豁免，主要原因是美国的稀土消费对中国的依存度很大。因此，
稀土行业成为了近几年特别热门的行业，与美国的贸易摩擦密切相关。
本文通过相关文献资料的梳理研究，对于稀土的概念、种类及作用有了清醒
的认识，以及对稀土行业发展现状以及未来发展趋势有了比较明确的了解，并对
如何选择股票估值方法有了明确的思路。通过对绝对估值法以及相对估值法在稀
土行业的适用性进行分析，从而选出合适的估值方法。在此基础上，通过对神经
网络模型进行学习、归纳、运用，并运用计算机软件进行建模，利用传统估值方
法并结合机器算法基于神经网络模型对稀土行业上市公司股票进行估值，进而能
够更具有创新性及实用性地对稀土行业上市公司进行股票估值研究，从而提供一
种估值方法，弥补目前研究的空白，最终能够一定程度地对稀土行业上市公司的
未来发展前景提供依据以及指导，进而能够进一步更好地发展和保护我国重要战
略资源，对后续的继续研究铺路。
关键词：中美贸易战，稀土行业，相对估值法，神经网络

青海大学硕士学位论文
Abstract
Abstract
Trade frictions between China and the United States have continued to heat up
since 2018. On July 10, 2018, the U.S. began to wave its trade bar, deciding to
increase tariffs on $200 billion of goods from China by a 10 percent increase, and
published a list of tariff-imposed goods in which rare earth resources can be included,
including rare earth oxides, compounds and other rare earth-related commodities.
However, two months later, on September 17th the United States continued to
increase its code, re-issued a tax increase list, decided to increase tariffs on another
$200bn of goods from China by 25 per cent, and on May 14th the following year the
United States issued a third tax increase list, deciding to tariff $300bn of goods,but
rare earths were exempted from both trade lists. As a result, the rare earth industry has
become a particularly popular industry for the past two years, closely related to trade
frictions with the United States.
Through the research of relevant literature, this paper has a clear understanding
of the concept, types and functions of rare earth, a clear understanding of the
development status and future development trend of rare earth industry, and a clear
idea on how to choose the stock valuation method. Through the analysis of the
applicability of absolute valuation method and relative valuation method in rare earth
industry, the appropriate valuation method is selected. On this basis, by studying,
summarizing and applying the neural network model, and using MATLAB machine
algorithm software, using traditional valuation methods and combining machine
algorithms based on neural network models to value listed companies in the rare earth
industry, it can be more innovative and practical to study the stock valuation of rare
earth industry listed companies, so as to provide a valuation method to make up for
the current research gap, and finally to provide a basis and guidance for the future
development prospects of listed companies in the rare earth industry, which in turn
can further and better develop and protect China's important strategic resources,
paving the way for further research.
Keywords: Sino-US trade war Rare earth industry Relative valuation Neural Network
。。。以下略