However, the effectiveness of semi-supervised anomaly detection hinges on the accuracy of the ’normal’ data labeling and may miss anomalies that subtly blend with the normal patterns . In semi-supervised anomaly detection, collecting data is generally easier compared to supervised methods, as it primarily involves identifying periods or regions in the time series where data is predominantly or entirely normal, thus reducing the need for extensive manual labeling. Anomalies are significantly rarer than normal data: This assumption underpins the effectiveness of various algorithms that identify anomalies as significant deviations from the majority of the data.