Description
The specified value of the Data Available per Tree (%) parameter results in fewer than six data values per tree. At least six data values per tree are required for regression.
Solution
Increase the value of the Data Available per Tree (%) parameter so that there are at least six data values for each tree. Ten percent of the input features (rounded up) are assumed to be withheld for validation. Multiply the remaining number of features by the percent of data per tree (rounded down) to determine how many data values will be available for each tree.
For example, you have 105 features and have 6 percent of the data available per tree. In this case, 11 values are assumed to be withheld for validation (105 * 10/100 = 10.5, which rounds up to 11), leaving 94 features available for training. Each tree will only have five data values, which is less than the required minimum of six (94 * 6/100 = 5.64, which rounds down to 5).