|
|
static Distribution | getDistribution (String str) |
| |
| static ContinuousDistribution | getDistributionMLE (String distName, double[] x, int n) |
| | Uses the Java Reflection API to construct a ContinuousDistribution , DiscreteDistributionInt or DiscreteDistribution object, as specified by the string str. More...
|
| |
| static DiscreteDistributionInt | getDistributionMLE (String distName, int[] x, int n) |
| | Uses the Java Reflection API to construct a DiscreteDistributionInt object by estimating parameters of the distribution using the maximum likelihood method based on the \(n\) observations in table \(x[i]\), \(i = 0, 1, …, n-1\). More...
|
| |
| static< T extends ContinuousDistribution > T | getDistributionMLE (Class< T > distClass, double[] x, int n) |
| | Uses the Java Reflection API to construct a ContinuousDistribution object by estimating parameters of the distribution using the maximum likelihood method based on the \(n\) observations in table \(x[i]\), \(i = 0, 1, …, n-1\). More...
|
| |
| static< T extends DiscreteDistributionInt > T | getDistributionMLE (Class< T > distClass, int[] x, int n) |
| | Uses the Java Reflection API to construct a DiscreteDistributionInt object by estimating parameters of the distribution using the maximum likelihood method based on the \(n\) observations in table \(x[i]\), \(i = 0, 1, …, n-1\). More...
|
| |
| static ContinuousDistribution | getContinuousDistribution (String str) |
| | Uses the Java Reflection API to construct a ContinuousDistribution object by executing the code contained in the string str. More...
|
| |
| static DiscreteDistribution | getDiscreteDistribution (String str) |
| | Same as getContinuousDistribution, but for discrete distributions over the real numbers. More...
|
| |
| static DiscreteDistributionInt | getDiscreteDistributionInt (String str) |
| | Same as getContinuousDistribution, but for discrete distributions over the integers. More...
|
| |
This class implements a string API for the package probdist.
It uses Java Reflection to allow the creation of probability distribution objects from a string. This permits one to obtain distribution specifications from a file or dynamically from user input during program execution. This string API is similar to that of UNURAN [166] .
The (static) methods of this class invoke the constructor specified in the string. For example,
d = DistributionFactory.getContinuousDistribution ("NormalDist (0.0,
2.5)");
is equivalent to
d = NormalDist (0.0, 2.5);
The string that specifies the distribution (i.e., the formal parameter str of the methods) must be a valid call of the constructor of a class that extends ContinuousDistribution or DiscreteDistribution, and all parameter values must be numerical values (variable names are not allowed). If no parentheses follow, the default parameters are used, i.e. a no-parameter constructor is searched and invoked. Distribution parameters are surrounded with parentheses and are separated from each other using commas. The order and types of the parameters should be the same as the corresponding class constructor. When specifying decimal values, the dot should be used for decimal separation, not the comma. When specifying a float or a double, one can use -infinity or infinity to denote \(-\infty\) and \(+\infty\), respectively. This is parsed to Double.NEGATIVE_INFINITY and Double.POSITIVE_INFINITY, respectively. However, this is not accepted by all probability distributions.
For example, if one uses the string Normal, a NormalDist object with \(\mu= 0\) and \(\sigma= 1\) will be created. If one uses Exponential(2.5), an ExponentialDist object with \(\lambda=2.5\) will be constructed.
The distribution parameters can also be estimated from a set of observations instead of being passed to the constructor. In that case, one passes the vector of observations, and the constructor estimates the parameters by the maximum likelihood method.
