Developer Guide for Intel® Data Analytics Acceleration Library 2016 Update 4
Most of algorithms in Intel DAAL have a set of algorithm-specific parameters. Because most of the parameters are optional and preset with default values, you can often omit parameter modification. Provide required parameters to the algorithm using the constructor during algorithm initialization. If you need to change the parameters, you can do it by accessing the public field parameter of the algorithm. Some algorithms have an initialization procedure that sets or precomputes specific parameters needed to compute the algorithm. You can use the InitializationProcedureIface interface class to implement your own initialization procedure when the default implementation does not meet your specific needs.
Each algorithm also has generic parameters, such as the floating-point type, computation method, and computation step for the distributed processing mode. In C++, these parameters are defined as template parameters, and in most cases they are preset with default values. You can change the template parameters while declaring the algorithm. In Java, the generic parameters have no default values, and you need to define them in the constructor during algorithm initialization.
For a list of algorithm parameters, refer to the description of an appropriate algorithm.