Pregled bibliografske jedinice broj: 1163858
Anomaly Correction in Time Series Data for Improved Forecasting
Anomaly Correction in Time Series Data for Improved Forecasting // 16th International Conference on Telecommunications (ConTEL 2021)
Zagreb, Hrvatska: Institute of Electrical and Electronics Engineers (IEEE), 2021. 21073867, 4 doi:10.23919/contel52528.2021.9495986 (predavanje, međunarodna recenzija, cjeloviti rad (in extenso), ostalo)
CROSBI ID: 1163858 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Anomaly Correction in Time Series Data for Improved
Forecasting
Autori
Ostroski, Dominik ; Slovenec, Karlo ; Brajdic, Ivona ; Mikuc, Miljenko
Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), ostalo
ISBN
978-9-5318-4271-6
Skup
16th International Conference on Telecommunications (ConTEL 2021)
Mjesto i datum
Zagreb, Hrvatska, 30.06.2021. - 02.07.2021
Vrsta sudjelovanja
Predavanje
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
time series , anomaly detection , time series fore-casting
Sažetak
This paper presents a method for detecting and correcting anomalies in time series data. This method was tested on time series data of disk usage over a period of few months. For the method to be able to detect and correct anomalies, it has to calculate the difference of time series, find the mean value of transformed data and use it to set a threshold. Any point in transformed data that has a value higher than the threshold corresponds to an anomaly in original data. After an anomaly is found, data is transformed in such a way that all data before the anomaly is shifted by the value of the anomaly. By removing anomalies this way, trend and seasonality of time series are kept intact. Results show that time series forecasting performed on transformed disk usage time series produces better results than when the original data is used.
Izvorni jezik
Engleski
