- ValueColumn
- Specifies the name of the input table column that contains the time series data.
- Accumulate
- Specifies the names of the input table columns to copy to the output table.Tip: To identify change points in the output table, specify the columns that appear in partition_exp and order_by_exp.
- SegmentationMethod
- [Optional] Specifies the segmentation method:
-
'normal_distribution' (Default)
In each segment, the data is in a normal distribution.
-
'linear_regression'
In each segment, the data is in linear regression.
-
'normal_distribution' (Default)
- SearchMethod
- [Optional] Specifies the search method, binary segmentation.
- MaxChangeNum
- [Optional] Specifies the maximum number of change points to detect. Default: 10.
- Penalty
- [Optional] Specifies the penalty function, which is used to avoid over-fitting:
-
'BIC' (Default)
The condition for the existence of a change point is:
ln(L 1)−ln(L 0) > (p 1 -p 0)*ln(n)/2
For normal distribution and linear regression, the condition is:
(p 1 -p 0)*ln(n)/2 = ln(n)
-
'AIC'
The condition for the existence of a change point is:
ln(L 1)−ln(L 0) > p 1 -p 0
For normal distribution and linear regression, the condition is:
p 1 -p 0 = 2
-
threshold, a DOUBLE PRECISION value:
The function compares the specified value to ln(L 1)−ln(L 0).
L 1 and L 2 are the maximum likelihood estimation of hypotheses H 1 and H 0. For normal distribution, the definition of Log(L 1 ) and Log(L 0) are in Background.
p is the number of additional parameters introduced by adding a change point. p is used in the information criterion BIC or AIC. p 1 and p 0 represent this parameter in hypotheses H 1 and H 0 separately.
-
'BIC' (Default)
- OutputOption
- [Optional] Specifies the output table columns. See the Output section. Default: 'CHANGEPOINT'.