Anomaly detection for short time series data in waste management

abstract

en

Anomaly detection is a very important step in every analysis of real-world data. Presence of the anomalies may strongly affect results of both tested hypotheses and created models. Data analysis is important in waste management to improve effective planning from both short- and long-term perspective. However, in the field of waste management, anomaly detection is rarely done. The goal of our paper is to propose a complex framework for anomaly detection in a big number of short time series. In such a case, it is not possible to use only an expert-based approach due to the time-consuming nature of this process and subjectivity. Proposed framework consists of two steps: 1. outlier detection via outlier test for trend adjusted data, 2. changepoints (trend changepoint, step changepoint) are identified via comparison of linear model parameters. Proposed framework is demonstrated on waste management data from the Czech Republic.

Anomaly detection is a very important step in every analysis of real-world data. Presence of the anomalies may strongly affect results of both tested hypotheses and created models. Data analysis is important in waste management to improve effective planning from both short- and long-term perspective. However, in the field of waste management, anomaly detection is rarely done. The goal of our paper is to propose a complex framework for anomaly detection in a big number of short time series. In such a case, it is not possible to use only an expert-based approach due to the time-consuming nature of this process and subjectivity. Proposed framework consists of two steps: 1. outlier detection via outlier test for trend adjusted data, 2. changepoints (trend changepoint, step changepoint) are identified via comparison of linear model parameters. Proposed framework is demonstrated on waste management data from the Czech Republic.

Waste management; short time series; anomaly detection; outlier; trend changepoint; step changepoint

17.10.2021

1847-7178