摘要 |
<P>PROBLEM TO BE SOLVED: To provide an anomaly analysis device to determine the type of anomalies based on inputted data. <P>SOLUTION: A training data input unit 101 receives input of training data and pre-defined parameters from the outside of the anomaly analysis device. A parameter estimation unit 102 estimates parameters for a linear combination model, based on the inputted training data. An evaluation data input unit 103 estimates parameters for a linear regression analysis model, based on the inputted evaluation data. An inter-series anomaly detection unit 104 detects whether or not an inter-series anomaly has been caused based on the estimated parameters for the linear combination model and the linear regression analysis model. An intra-series anomaly detection unit 105 detects whether or not an intra-series anomaly has been caused based on the estimated parameters for the linear combination model and the linear regression analysis model. <P>COPYRIGHT: (C)2012,JPO&INPIT |