Bounded Mean Change Segmentation

How to use Bounded Mean Change Segmentation

Keep observations in their supplied order, enter a minimum segment length, then choose an exact count or a maximum with a user-selected penalty. Inspect all best-by-count costs and the complete deterministic break witness.

1. Use the example as the schema; replace its labels and entered model assumptions.
2. Keep observations in their supplied order, enter a minimum segment length, then choose an exact count or a maximum with a user-selected penalty. Inspect all best-by-count costs and the complete deterministic break witness.
3. Calculate, inspect the complete witnesses and limitations, then copy or download the JSON report.

Good to know

• 1–120 exact decimal-string observations within ±1,000,000,000, with at most six places. Minimum segment length is 1–120. Fixed count or maximum count is 1–10; every selected segment must meet that length. Missing positions and decimal JSON numbers are refused.
• Exact integer-scaled prefix sums avoid large-offset cancellation. The bounded dynamic program examines all permitted contiguous segment endpoints; it is not a streaming or PELT implementation. Output means and costs remain fractions.
• The penalty is an entered nonnegative quantity up to 1,000,000,000, in squared observation units. Equal penalized costs prefer fewer segments; equal fixed-count costs prefer the earliest complete boundary vector. No statistical discovery guarantee, p-values, causal claim, automatic penalty choice or forecast.
• Local calculation with no upload, input persistence or outside lookup. Model JSON is limited to 30,000 characters and 10,000 values. Decimal quantities are strings with up to six places; count fields use plain JSON integers. Exact fractions are capped at 8,192 bits. The complete JSON output is limited to 1,000,000 characters; approximate numbers preserve their stored values. Unsupported states fail explicitly.