Manipulation in Prediction Markets - Chasing the Fraudsters
Manipulation in Prediction Markets - Chasing the Fraudsters
复制标题
预测市场的操纵——追捕欺诈者
DOI：
--
复制
发表时间：
2017
期刊：
影响因子：
--
通讯作者：
Tobias T. Kranz
中科院分区：
文献类型：
--
作者：
Simon Kloker;Tobias T. Kranz
预测市场是预测和企业知识管理的常用工具。基于“群众的智慧”，它的预测经常超过民意调查和统计模型。此外，它还提供了一种方便的方式来收集组织中分散的信息，并鼓励员工透露私人信息并保持知情。虽然这类市场已经建立，但在其运作和维护方面仍然存在一些悬而未决的问题。特别是在许多情况下报告的操纵和欺诈问题，很少得到解决;如果是这样，也只是非常理论化或使用复杂的算法，难以为从业者实现。然而，缺乏一个严格的框架，揭示预测市场的弱点，并提供适用的预防和检测策略。我们提出了欺诈立方体，一个简洁的框架揭示欺诈者的思维过程，从而潜在的攻击向量。此外，我们提出了一个易于实现的检测算法的基础上，最先进的检测算法。最后，我们显示不低于可比的检测率建立检测算法，同时提供上级适用性。
Prediction markets are a common instrument in forecasting and corporate knowledge management. Based on the “wisdom of the crowd” its forecasts regularly outperform polls as well as statistical models. In addition, it offers a convenient way to collect dispersed information in organizations and incite employees to reveal private information as well as to stay informed. Although such markets are well established, there still remain open questions regarding their operation and maintenance. Especially the issue of manipulation and fraud, which are reported in many cases, is only rarely addressed; if so, only very theoretical or with complex algorithms, hard to implement for practitioners. Yet, a rigid framework, uncovering weaknesses of prediction markets and offering applicable prevention and detection strategies is missing. We propose the Fraud Cube, a concise framework unveiling fraudster’s thought process and thus potential attack vectors. Additionally, we present an easy to implement detection algorithm based on state of the art detection heuristics. Finally, we show not less than comparable detection rates to established detection algorithms whilst providing superior applicability.