随着互联网技术的普及，移动社交媒体逐渐步入人们的生活，但各类复杂的信息噪声也增加了人们筛选有效信息的成本。目前具有代表性的降噪方法是评估用户的影响力，但目前国内外在垂直领域方面研究的影响力模型较少。因此，本文选取了雪球网这一成熟的投资领域社会化平台，探究在该领域如何建立用户影响力评估模型，以丰富用户影响力相关的学术研究，同时也为社交平台的投资者提供一个低成本、高效率的信息筛选环境。通过借鉴、分析不同平台的影响力模型，结合数据特征，本文选择了主成分分析法，以确定转发数、粉丝数等10个指标的权重。通过各主成分及方差贡献率计算各指标在主成分线性组合中的系数，再进行加权平均和归一化计算，得到了最终的影响力模型。为了验证其有效性，本文采取了层次分析法和和熵值法，再次计算影响力并进行对比。实验结果显示，本文所提出的用户影响力评估方法，能够有效地适用于投资领域的社交平台，准确呈现不同用户的影响力。
With the popularity of Internet technology, mobile social media has gradually entered people’s lives, but at the same time all kinds of complex information noise has increased the cost of screening effective information. At present, the representative noise reduction method is to evaluate the influence of users, but few studies on influence models in the vertical field at home and abroad. Therefore, this paper explores how to establish a user impact assessment model by analyzing Snowball Network, a mature investment socialization platform, so as to enrich the academic research in this field and also provide a low-cost, high-efficiency information screening environment for investors in social platforms. Combining data characteristics, this paper selects the PCA method to determine the weights of 10 indicators such as the number of forwarding, fans, portfolio concerns. The coefficients of each index in the linear combination of principal components are calculated by the principal component and variance contribution rate, and then the weighted average and normalized calculations are performed to obtain the final influence model. In order to verify its validity, the AHP and entropy method were used to calculate the influence again. The experimental results show that the user impact assessment method proposed in this paper can be effectively applied to social platform of investment field and accurately present the influence of different users.
