﻿ 基于洛伦兹变换和PageRank算法的数据资产估值
 计算机系统应用  2020, Vol. 29 Issue (8): 205-210 PDF

1. 云南机电职业技术学院 工业信息技术系, 昆明 650704;
2. 云南财经大学 云南省经济社会大数据研究院, 昆明 650221

Data Asset Valuation Based on Lorentz Transform and PageRank Algorithm
SUN Xiao-Xuan1, ZHAO Xiao-Ming2
1. Department of Industrial Information Technology, Yunnan Vocational College of Mechanial & Electrial Technology, Kunming 650704, China;
2. Big Data Research Institute of Yunnan Economy and Society, Yunnan University of Finance and Economics, Kunming 650221, China
Foundation item: Science and Technology Innovation Team for Business Intelligence of Higher Educations of Yunnan Province (42212217010)
Abstract: Data resources are important production materials that make up a digital society, and the evaluation of the value of data resources is an important basis for data transactions, data circulation, and data appreciation. Based on the theoretical basis of the Lorentz transform and PageRank algorithm, firstly, we use the PageRank algorithm to calculate the weight coefficient in the data asset pricing system, and get the initial valuation of the data asset. Then, we use the data asset valuation model of the quality-speed relationship mapping to value the given data asset. The experimental results show that the proposed data asset valuation method is of certain efficiency and market reference utility.
Key words: Lorentz transformation     PageRank     data resource     data assets     asset valuation

 $m = \dfrac{{m_0}}{{\sqrt {1 - {{\left( {\dfrac{v}{c}} \right)}^2}} }}$ (1)

1 问题陈述

 $D\left( {{d_{a0}},r} \right) = {d_{a0}} \times \left( {1 + r} \right)$ (2)

(1)对于D, 如在狭义相对论中洛伦兹变换, 描述参照第1节中式(1)所示. 我们将数据资产估值与物体的惯性质量进行类比, 将数据资产初始估计与原本静质量进行类比, 数据资产交易频率与物体速度进行类比. 因此, 本文以洛伦兹变换作为研究数据资产估值的计量方法的支撑理论基础, 在本文第2节将就此重点展开.

 ${D_a} = \dfrac{{{d_{a0}}}}{{\sqrt {1 - {{\left( {\dfrac{{c - v}}{c}} \right)}^2}} }}\left( {1 + r} \right)$ (3)

 ${d_{a0}} = \dfrac{{\left( {\displaystyle\sum\limits_{i = 1}^n {{p_i} \times {w_i}} } \right)}}{n}$ (4)

wi: 修正系数. $n \ge 3$ , 具体可参照数据资产的数量不少于3个.

(2)对于r, 我们采用PageRank算法来衡量数据资产估值的重要程度, 并计算得到数据资产的估值系数. 我们以PageRank算法作为理论依据, 提出如下2点假设:

 $R({d_a}) = {\left( {(1 - d) + d\sum\limits_{i = 1}^n {\dfrac{{P({d_i}) \times {U_i}}}{{C({d_i})}}} } \right)^*}$ (5)

2 数据资产估值

2.1 数据资产的初始估值计量

(1)数据资产特征维度

 ${{W}}_i = \dfrac{1}{n} - \dfrac{1}{{2a}} + \dfrac{1}{{na}}\sum\limits_{j = 1}^n {a_{ij}};\;\;i = 1,2,\cdots,n,\;a \ge \dfrac{{n - 1}}{2}$ (6)

 ${W}_{{\rm{1 - 9}}} = \left( {0.05,0.11,0.15,0.25,0.14,0.11,0.1,0.04,0.05} \right)$

(2)数据资产估值分析模型

 ${d_{a0}} = \dfrac{{\left( {\displaystyle\sum\limits_{i = 1}^n {{p_i} \times {{\left( {\sum {\left( {\alpha + \beta + \gamma + \delta + \lambda } \right)} } \right)}_i}} } \right)}}{n}$ (7)

(3)计算数据资产初始估值

① 假设有一数据资产i需上市交易, 其初始估值设为di.

② 通过网络抓取技术, 建立数据资产i的可比数据资产价值结构化实例库, 然后通过专家打分法确定特征维度的权重系数. 为避免专家打分过强主观, 实验采用2组专家分别交叉打分, 并对结果进行一致性检验. 经Kappa检验, 得到Kappa值为0.763, 说明对于特征的权重一致性较好. 对于特征权重取值为分组打分的均值, 结果如表3所示.

③ 合成初步估值结果

 $\begin{split} da0 = &(\left( {1.00 \times \left( {0.2 + 0.2 + 0.18 + 0.15 + 0.2} \right)} \right) +\\ & 0.85 \times \left( {0.2 + 0.2 + 0.19 + 0.2 + 0.2} \right)+\\ & \left( {0.78 \times \left( {0.2 + 0.2 + 0.2 + 0.2 + 0.15} \right)} \right)/3 = 0.84 \end{split}$
2.2 数据资产重要程度估值系数

 $\begin{split} R(it) = & ( 1 - 0.85) + 0.85\times(( 1/5\times0.8 ) + ( 0.85/8\times0.5 ) +\\ & (0.78/6\times0.5)) = 0.39 \end{split}$
2.3 数据资产状态估值

 $D\left( i \right) = 0.84/{\rm{SQRT}}\left( {1 - {{\left( {\left( {30 - 7} \right)/30} \right)}^2}} \right)\times\left( {1 + 0.39} \right) = 1.85$
3 实验分析

 图 1 随着数据资产数量增加的执行时间变化趋势

 图 2 数据资产实验计算估值与人工估值比对图

 图 3 数据资产实验计算估值与市场价格比对图

4 总结展望

 [1] 徐漪. 大数据的资产属性与价值评估. 产业与科技论坛, 2017, 16(2): 97-99. DOI:10.3969/j.issn.1673-5641.2017.02.051 [2] 邹照菊. 关于大数据资产计价的若干思考. 财会通讯, 2018(28): 35-39. DOI:10.16144/j.cnki.issn1002-8072.2018.28.009 [3] Kozyr YV. Business valuation based on assets replacement cost. Journal of Finance and Investment Analysis, 2014, 3(3): 1-12. [4] Lagrost C, Martin D, Dubois C, et al. Intellectual property valuation: How to approach the selection of an appropriate valuation method. Journal of Intellectual Capital, 2010, 11(4): 481-503. DOI:10.1108/14691931011085641 [5] Gao Y, Lim KG, Ng KH. An approximation pricing algorithm in an incomplete market: A differential geometric approach. Finance and Stochastics, 2014, 8(4): 501-523. [6] 张驰. 数据资产价值分析模型与交易体系研究[博士学位论文]. 北京: 北京交通大学, 2018. [7] 戴又善, 戴亮. 普遍的质速关系和狭义相对论. 北京大学学报(自然科学版), 2014, 50(3): 403-410. DOI:10.13209/j.0479-8023.2014.063 [8] Qin XP, Li P, Su G, et al. Derivation of Lorentz transformation via Lorentz invariant. Proceedings of the 2nd International Conference on Automatic Control and Information Engineering. Beijing, China. 2017. 50–53. [9] 张元仲. 狭义相对论洛伦兹变换的推导及其他. 物理与工程, 2016, 26(3): 3-8. DOI:10.3969/j.issn.1009-7104.2016.03.001 [10] 王嵘冰, 安维凯, 冯勇, 等. 基于标签和PageRank的重要微博用户推荐算法. 计算机科学, 2018, 45(2): 276-279. DOI:10.11896/j.issn.1002-137X.2018.02.047 [11] 赵小明, 张群, 岳昆. 基于静电场理论和PageRank算法的微博用户相关性分析. 云南大学学报(自然科学版), 2015, 37(2): 207-214. DOI:10.7540/j.ynu.20140430 [12] Lv LS, Zhang K, Zhang T, et al. PageRank centrality for temporal networks. Physics Letters A, 2019, 383(12): 1215-1222. DOI:10.1016/j.physleta.2019.01.041 [13] 李伟霖, 王成良, 文俊浩. 基于评论与评分的协同过滤算法. 计算机应用研究, 2017, 34(2): 361-364, 412. DOI:10.3969/j.issn.1001-3695.2017.02.009 [14] 呼亚杰. 一种基于类别偏好协同过滤推荐算法的实现与优化[硕士学位论文]. 兰州: 兰州大学, 2016.