ABSTRACT The diffusion of Internet-based Intangible Network Goods (IINGs) shows new characteristics completely different from that of traditional material products. This paper aims to establish new models to describe and predict IING’s diffusion at the aggregate level. Firstly, we transform the key factors affecting IING’s diffusion into driving forces, resistant forces, and variable forces. Secondly, we analyse the dynamic changes of these forces in different diffusion stages and obtain the acceleration model of IING’s diffusion. Then, since acceleration is the second derivative of scale, we further establish the scale model of IING’s diffusion. As the scale model can predict the number of IING’s adopters at a particular time and the acceleration model can explain the dynamic changes of scale, we combine them as the acceleration-scale model to describe IING’s diffusion. Finally, we make comparisons between the acceleration-scale model and the Bass model based on three cases. Different from the previous studies, we found that IING’s diffusion rate is asymmetric. The diffusion rate of successful IING is right skewed while the diffusion rate of failed IING is left skewed. The results also shows that the acceleration-scale model has a better predictive performance than the Bass model, no matter the diffusion is successful or failed

中文翻译：

---

基于力分析的IING扩散加速度尺度模型

摘要 基于互联网的无形网络商品（IINGs）的传播呈现出与传统物质产品完全不同的新特征。本文旨在建立新的模型来描述和预测 IING 在总体层面的扩散。首先，我们将影响IING扩散的关键因素转化为驱动力、阻力和可变力。其次，我们分析了这些力在不同扩散阶段的动态变化，得到了IING扩散的加速度模型。然后，由于加速度是尺度的二阶导数，我们进一步建立了IING扩散的尺度模型。由于规模模型可以预测特定时间 IING 的采用者数量，而加速度模型可以解释规模的动态变化，我们将它们组合为加速度尺度模型来描述 IING 的扩散。最后，我们基于三种情况对加速度尺度模型和Bass模型进行了比较。与以往的研究不同，我们发现 IING 的扩散速率是不对称的。成功的 IING 的扩散率是右偏的，而失败的 IING 的扩散率是左偏的。结果还表明，无论扩散是成功还是失败，加速度尺度模型都比Bass模型具有更好的预测性能 成功的 IING 的扩散率是右偏的，而失败的 IING 的扩散率是左偏的。结果还表明，无论扩散是成功还是失败，加速度尺度模型都比Bass模型具有更好的预测性能 成功的 IING 的扩散率是右偏的，而失败的 IING 的扩散率是左偏的。结果还表明，无论扩散是成功还是失败，加速度尺度模型都比Bass模型具有更好的预测性能