论文部分内容阅读
应用去趋势波动分析法,对电子商务中人类网上购物行为进行研究,首次探讨了人类浏览及购买行为时间序列(数量波动)标度律.首先,研究发现人类网上购物行为呈现出明显的周期性,其时间序列的概率密度函数具有显著的双模态特征.其次,利用傅里叶变换方法分析浏览以及购买行为时间序列的功率谱,发现其演化过程不同于无关联的泊松过程.最后,基于功率谱过滤周期性趋势的影响,对去除周期趋势后的浏览和购买行为时间序列进行去趋势波动分析,发现其标度行为表明其具有较强的长程关联特性,且平均标度值近似为1,表明其具有自组织临界性.实证研究结果与其他领域如因特网交通流和金融市场价格波动的标度行为相似,有助于理解人类活动如何影响电子商务系统演化和提高在线商务活动效率,对分析电子商务中人类行为活动的机制和预测其波动趋势具有重要的启示作用.
This paper studies the behavior of human online shopping in e-commerce by using the trend-less volatility analysis method, and first discusses the scale law of time series (quantity fluctuation) of human browsing and purchasing behavior.First, the study finds that human online shopping behavior shows obvious periodicity , The probability density function of time series has significant bimodal characteristics.Secondly, the power spectrum of browsing and purchasing behavior time series is analyzed by using Fourier transform method, and the evolution process is different from that of uncorrelated Poisson process.Finally, Based on the influence of the periodic trend of power spectrum filtering, the trend analysis of the browsing and purchasing behavior after the removal of the periodic trend is conducted to analyze the trend fluctuation. The scaling behavior shows that it has strong long-range correlation and the average scale value is approximately 1, which indicates that it has self-organized criticality.The empirical results are similar to those in other fields such as Internet traffic flow and the price fluctuation of financial markets, which helps to understand how human activities affect the evolution of e-commerce systems and the efficiency of online business activities, It is important to analyze the mechanism of human behavior in e-commerce and forecast the trend of its fluctuation Inspiration.