A Robust Data-Driven Method for Multiseasonality and Heteroscedasticity in Time Series Preprocessing

<div>The proposed DOW and DOW-FLH gives similar AUC and around 27% bigger coverage than the others on average. Besides, DOW-FLH produces only half false positives compared to DOW without FLH.</div>

Wireless Communications and Mobile Computing

fig5

Figure 5

Figure 5: A Robust Data-Driven Method for Multiseasonality and Heteroscedasticity in Time Series Preprocessing 