Added exponential features generator for test data. Tweeked up the PerceptronSandbox and brought the LogisticRegressionSandbox in line with datasource api
added checks to make sure autoranges don't take on NaN or infinite values
Conflicts: main/math/src/main/java/org/rl337/math/types/Matrix.java main/skynet/src/main/java/org/rl337/skynet/types/BinaryThreshold.java main/skynet/src/main/java/org/rl337/skynet/types/Sigmoid.java
starting activation function stuff.
added probability density function which takes a range to calculate probability for. added the ability to sample the distribution.
…e that's negative infinity and infinity respectively. Mad normal distribution q function handle mu and sigma params
… normal distributions
…de in the interest objects
Removed extra sumColumns from hypothesis