Streaming data services, such as video-on-demand, are getting increasingly more popular, and they are expected to account for more than 80% of all Internet traffic in 2020. In this context, it is important for streaming service providers to detect deviations in service requests due to issues or changing end-user behaviors in order to ensure that end-users experience high quality in the provided service. Therefore, in this study we investigate to what extent sequence-based Markov models can be used for anomaly detection by means of the end-users’ control sequences in the video streams, i.e., event sequences such as play, pause, resume and stop. This anomaly detection approach is further investigated over three different temporal resolutions in the data, more specifically: 1 h, 1 day and 3 days. The proposed anomaly detection approach supports anomaly detection in ongoing streaming sessions as it recalculates the probability for a specific session to be anomalous for each new streaming control event that is received. Two experiments are used for measuring the potential of the approach, which gives promising results in terms of precision, recall, \(F_1\)-score and Jaccard index when compared to k-means clustering of the sessions.

中文翻译：

---

使用多个时间分辨率和马尔可夫链对事件序列进行异常检测

点播视频等流数据服务正变得越来越流行，预计到2020年，它们将占所有Internet流量的80％以上。在这种情况下，对于流服务提供商而言，检测偏差至关重要由于问题或最终用户行为发生变化而导致的服务请求中的错误，以确保最终用户体验所提供服务的高质量。因此，在这项研究中，我们研究了基于序列的马尔可夫模型在多大程度上可以通过视频流中最终用户的控制序列（即事件序列，如播放，暂停，恢复和停止）用于异常检测。在数据中的三种不同的时间分辨率上，进一步研究了这种异常检测方法，更具体地说：1 h，1天和3天。所提出的异常检测方法支持正在进行的流会话中的异常检测，因为它重新计算特定会话对于接收到的每个新流控制事件而言异常的概率。进行了两次实验以测量该方法的潜力，这些方法在精度，召回率，\（F_1 \）- score和Jaccard索引与会话的k -means聚类相比。