| Sturges |
Sturges algorithm performs best if data is normally distributed and n is at least 30.
Algorithm for calculating bin width:w = r/(1 + log2n)
where:
w = bin width
r = data value range
n = number of elements in dataset
|
| Scott |
Scott algorithm performs best on normally distributed data.
Algorithm for calculating bin width:w = 3.49s/(n1/3)
where:
w = bin width
s = standard deviation of data values
n = number of elements in dataset
r = data value range
Number of bins: r/w
|
| Variable-Width |
|
| Equal-Width |
Algorithm for calculating bin width: w = (max - min)/k
where:
min = minimum value of the bins
max = maximum value of the bins
k = number of intervals into which algorithm divides dataset
Interval boundaries: min+w, min+2w, …, min+(k-1)w
- Optional MinMax table.
- If MinMax table is omitted, the TD_Histogram function internally computes the min value and max value from the input data for the target columns.
- If MinMax table is specified, the user can specify in the following manner:
- If one target column is specified, specify min value in column1 and max value in column2.
- If more than one target column is specified, specify ColumnName in column1, min value in column2, and max value in column3.
|