Wolfram Research has released Neural Networks, which lets users test and explore neural network models “faster and easier than ever before” using the computational power of Mathematica, according to company spokesperson Ben Wilson.
The new Neural Networks application — designed for use with Mathematica 4 or later on Mac OS X, the traditional Mac operating system, Windows, Linux, Solaris, HP-UX, IRIX, AIX, Compaq Tru64 Unix, and compatible systems — provides a robust environment for modeling data structures. Artificial neural networks have revolutionized the way researchers solve complex, real-world problems in engineering, science, economics, and finance, according to Yezabel Dooley, applications product manager at Wolfram Research.
The application was created to give professionals and students the tools to train, visualize and validate neural network models. It supports a set of neural network structures and implements training algorithms while utilizing Mathematica’s number crunching and visualization capabilities, Dooley said. It also includes special functions to address typical problems in data analysis, such as function approximation, classification and detection, clustering, nonlinear time series and nonlinear system identification problems.
The suggested commercial price is US$895; academic and volume discounts are available. Details and international pricing can be found at the Neural Networks product site.