Dynamic Threshold (DT)

YAML type: dt | Category: Adaptive

Flags values outside a threshold derived from recent history (adaptive baseline).

Description

DT builds an adaptive baseline from historyDays with configurable sensitivity and optional seasonality filtering of comparison days.

How it works

Dynamic Threshold (DT) illustration

Dynamic Threshold builds an adaptive baseline from recent historyDays, optionally filtering comparison days by seasonality (weekly, monthly, or annual).

The current value is checked against a sensitivity-scaled band around that baseline. Points outside the band become DT anomalies -- a general-purpose alternative to fixed percentual thresholds.

Setup

Pick seasonality (None, Weekly, Monthly, Annual) to match the series pattern.

Configure detectors on the data view Analysis tab or in YAML. Validate with Data Preview before enabling production schedules.

Settings

Parameter	Default	Description
`historyDays`	`60`	History for the adaptive baseline
`sensitivity`	`3.0`	Threshold sensitivity
`seasonality`	`None`	Seasonal day filter -- see Allowed values

Allowed values

`seasonality`

Controls which historical days contribute to the adaptive baseline (case-insensitive).

Value	Meaning
`None` (default)	No seasonal day filter -- use recent history as-is
`Weekly`	Prefer same weekday context
`Monthly`	Prefer same day-of-month context
`Annual`	Prefer same calendar-day-of-year context

Any unrecognized value is treated as None.

Shared series filters

Most detectors also accept optional series filters in params:

Parameter	Purpose
`maxGapAmount`	Skip series when gap amount is too high (`null` = no limit). Catalog create-default is often `5`.
`minValue`	Skip series whose average value is below this (`null` = no limit). Catalog create-default is often `1000`.
`maxValue`	Skip series whose average value is above this (optional; omit/`null` = no limit)
`useSpecialDates`	When `true`, suppress anomalies that overlap special dates

When a filter skips a series, Data Preview shows the reason.

Configuration example

analysis:
  detectors:
    - type: dt
      params:
        historyDays: 60
        sensitivity: 3.0
        seasonality: Weekly

Using the detector

Good general-purpose adaptive detector when you do not want fixed % thresholds.

Use cases

Series without strong business calendar rules
Mixed portfolio metrics with different scales
Continuous monitoring with auto-updating baselines

Specifics

For explicit seasonality residual modeling see sr.