Soft computing differs from conventional or hard computing or rule based computation. Unlike hard computing, soft computing programs are tolerant of imprecision and uncertainty in the final answer in exchange for the ability to solve more complex real world problems in an acceptable time frame.
We use signal processing, statistical analysis, combinatorics, genetic algorithms, neural networks and principle component analysis to extract meaningful information and predictability from what appears to be an intractable problem.