Method softmax

JSAT/src/jsat/math/MathTricks.java:91–102

Applies the softmax function to the given array of values, normalizing them so that each value is equal to exp(x j ) / Σ ∀ i exp(x i ) @param x the array of values @param implicitExtra true if the softmax will assume there is an extra implici

(double[] x, boolean implicitExtra)

     * an extra implicit value not included in the array with a value of 0.0 
     */
    public static void softmax(double[] x, boolean implicitExtra)
    {
        double max = implicitExtra ? 1 : Double.NEGATIVE_INFINITY;
        for(int i = 0; i < x.length; i++)
            max = max(max, x[i]);
        
        double z =implicitExtra ? exp(-max) : 0;
        for (int c = 0; c < x.length; c++)
            z += (x[c] = exp(x[c] - max));
        for (int c = 0; c < x.length; c++)
            x[c] /= z;
    }
    
    /**
     * Applies the softmax function to the given array of values, normalizing