JSMILE Learning: Dataset

This class holds data that can be used for learning. All learning algorithms take a data set as input.

Constants

• static final float DefaultMissingFloat

The default value for a missing value in case of a continuous variable.

• static final int DefaultMissingInt

The default value for a missing value in case of a discrete variable.

Methods

• int getRecordCount()

The number of records.

• int getVariableCount()

The number of variables.

• void addEmptyRecord()

Adds an empty record to the data set. The record will contain missing value markers.

• void addFloatVariable(String id)

Adds a continuous variable to the data set. This method used DefaultMissingFloat as missing value marker.

• void addFloatVariable(String id, float missingValue)

Adds a continuous variable to the data set and specify the value that indicates a missing value.

• void addIntVariable(String id)

Adds a discrete variable to the data set. This method used DefaultMissingInt as missing value marker.

• void addIntVariable(String id, int missingValue)

Adds a discrete variable to the data set and specify the value that indicates a missing value.

• float getFloat(int variable, int record)

Returns the corresponding continuous element in the data set. This method throws a SmileException when an invalid variable or record index is passed.

• int getInt(int variable, int record)

Returns the corresponding discrete element in the data set. This method throws a SmileException when an invalid variable or record index is passed.

• int getNodeHandle(int variable)

Returns the handle of the variable. This method throws a SmileException when an invalid variable index is passed.

• String[] getStateNames(int variable)

Returns the state names of the variable. This method throws a SmileException when an invalid variable index is passed.

• String getVariableId(int variable)

Returns the id of the variable. This method throws a SmileException when an invalid variable index is passed.

• void setFloat(int variable, int record, float value)

Sets the corresponding continuous element to a new value. This method throws a SmileException when an invalid variable or record index is passed.

• void setInt(int variable, int record, int value)

Sets the corresponding discrete element to a new value. This method throws a SmileException when an invalid variable or record index is passed.

• void setStateNames(int variable, String[] names)

Sets the state names of a variable. This method throws a SmileException when an invalid variable index is passed.

• double[] discretize(int variable, int algorithm, int intervals, String statePrefix);

The arguments are:

1) variable: 0-based column index

2) algorithm: algorithm IDs are defined in DataSet.DiscretizationAlgorithmType

3) intervals: number of discretization bins

4) statePrefix: prefix for the generated state names; pass null or empty string to skip state name generation

The return value is the array containing the bin edges, the length of the returned array is intervals - 1.