new SymbolicRegression(data)
Fitness function for fitting data, i.e. supervised learning. Symbolic regression is a type of
regression analysis that searches the space of mathematical expressions to
find the model that best fits a given dataset.
Evaluate fitness based on fitness cases and target values. Fitness cases are
a set of exemplars (input and output points) by comparing the error between
the output of an individual(symbolic expression) and the target values.
Parameters:
| Name | Type | Description |
|---|---|---|
data |
array | Exemplars to evaluate individual solutions on. The last column is the label |
- Source:
Methods
evaluate(node, inputData) → {number}
Evaluate the model (f) on the input data (x_0,...,x_n), i.e. f(x_0,...,x_n).
Evaluate a node recursively. The node's symbol is evaluated.
Parameters:
| Name | Type | Description |
|---|---|---|
node |
TreeNode | Node that is evaluated |
inputData |
array | Data of the input variables x_0, ..., x_n |
- Source:
Returns:
Value of the evaluation, i.e. f(x)
- Type
- number
evaluateIndividual(individual) → {number}
Evaluate the individual solution (model) on all the data. Fitness is the
negative Mean Squared Error.
Parameters:
| Name | Type | Description |
|---|---|---|
individual |
Individual | Individual solution to evaluate |
- Source:
Returns:
Fitness, the negative MSE
- Type
- number
evaluatePopulation(Population) → {array}
Assign fitness to each individual solution in the population.
Parameters:
| Name | Type | Description |
|---|---|---|
Population |
array | List of Individuals |
- Source:
Returns:
Evaluated population
- Type
- array