cVision Capabilities

cVision offers a broad range of flexibilities concerning data selection, neural network design and training, and of course quality control. One of its superior assets are the fully automatic and comprehensive network design and training features and the option to monitor the whole training process in a great variety of possibilities. cVision supports in a wide range the data selection, trains and grows an arbitrary number of networks automatically to find the optimal network size and stores the optimal network for further use. cVision offers

• Partial selection of data from multiple sheets
• Simple dedication of data sub-sets to different training tasks
• Handling of numerical, categorical and missing values
• Automatic network design and training
• Prevention from overtraining and over fitting
• Decision probabilities and confusion matrices

All with one mouse click, comprehensive and reproducible. cVision requires no prior knowledge of neural networks technology, with only minimal user interaction it addresses all the issues associated to building robust models from available real world data. cVision is not only fast and accurate, its overriding characteristic is the fully automatic network design and training feature to get the optimal network tailored to the particular problem, automatically determining the optimal network size. Prevention from overtraining and oversizing neural networks is standard in cVision.

cVision partitions the input data into proper subsets for learning, validation and testing, and then constructs, trains and stores a neural network tailored to the particular problem. For advanced users, cVision also offers direct access to all key training, network and data parameters.

To look at some cVision screen shots click here.