13: Industrie 4.0, Vorlesung, WS 2018/19, 25.01.2019
Torsten Kröger · 2019
13 | 0:00:00 Start 0:00:07 Syllabus - Winter term 2018/19 0:01:48 Literature for Today's class 0:02:41 Video: Simple Outlier Detection 0:06:25 Motivation: Predictive Maintenance 0:09:51 Machine Learning Algorithms for Preditictive Maintenance 0:16:27 Anomaly Detection 0:18:26 Related Problems 0:20:18 Relationship Among Data Instances 0:21:21 Rule-based Anomaly Detection 0:29:01 Representation Matters 0:31:02 Anomalies in a Single Time Series Signal 0:32:28 Anomalies in a Multiple Time Series Signals 0:34:20 Criteria for Anomaly Detection 0:39:42 Characteristics of Univariate, Multivariate and Hybrid Anomaly Detection Methods 0:49:31 Characteristics of Univariate, Multivariate and Hybrid Anomaly Detection Methods 0:54:12 Type of Anomalies 0:55:23 Contextual Anomalies 0:57:13 Collective Anomalies 0:58:42 Output of Anomaly Detection 0:59:22 Evaluation of Anomaly Detection - F-value 1:01:33 Accuracy of Anomaly Detection - F-value 1:03:26 Evaluation of Outlier Detection - ROC & AUC 1:07:04 General Scheme for Anomaly Detection 1:08:48 Taxonomy 1:12:11 Variants of Anomaly Detection Problem