NeuCom - A Neuro-computing Decision Support Enviroment
NeuCom represents a new generation of computer environment, it is self-programmable, learning and reasoning computer environment based on connectionist (Neurocomputing) modules. NeuCom learns from data, thus evolving new connectionist modules. The modules can adapt to new incoming data in an on-line incremental, life-long learning mode, and can extract meaningful rules that would help people discover new knowledge in their respective fields. NeuCom is based on the theory of Evolving Connectionist Systems (ECOS) as published in (“Evolving connectionist systems: methods and applications in bioinformatics, brain study and intelligent machines”, N.Kasabov, Springer Verlag, London, 2002)
NeuCom can be used to solve complex problems
Such problems are clustering, classification, prediction, adaptive control, data mining and pattern discovery from databases in a multidimensional, dynamic and possibly changing data environment. Applications span in all areas of Science, Engineering, Medicine, Bio-informatics, Business, Arts and Design, Education.
NeuCom is both a decision support system and a DSS development environment
NeuCom can be used either as a decisions support system (DSS), where users specify their task and define data to be used, in order to obtain a solution, or - as a DSS development environment for building sophisticated problem oriented intelligent DSS. The end users in the former case are people who have never programmed computers, but have databases available and need a decision to be made based on existing data and/or human knowledge. In the latter case users are professional system developers who can develop DSS for various applications in collaboration with experts in the field.
- The correlation coefficient module GUI changed, integrated p-values and r-values selection into a single interface, which should makes it more usable.
- 3D visualisation is updated to provide better colour for data with less than 8 classes. For classification problem with less than 8 classes, a fix set of most contrasting colours are used. For problem with large number of classes, the colours are generated automatically across the colour range.
- KNN for prediction in cross validation module is now working.
- Bi-Clustering module now included in NeuCom Student with updated colormap to improve visibility
- PCA module now has more configuration options for better visualisation.
- User interface issue on Windows Vista has been resolved. Cross Validation module should work correctly now.
- Fixed the problem with Signal to Noise Ratio (SNR) module.
- Rule display in ECOS modules are fixed
- Result display in cross validation module is fixed.
- ECF incremental learning has been implemented.
- ECF module's own cross validation is now working correctly.
- EFUNN progress bar display in incremental learning mode has been improved. It should no longer flicker.
- Text file loading is no longer repeated.
- DENFIS incremental mode added
- EFUNN module implemented include all standard options and rule extraction.
- Class Distribution plot module added, allows better visualisation of classification data.
- Cross Validation module added to allow various experiment setups with options for on the fly feature selection using signal to noise ratio.
- Data loading for Comma separated values (CSV) has been re-written to allow faster loading time and better error handling.
- Data size limitation and the cross validation module has been removed. This allows students to load data file of any size, limited only by the PC's hardware capability.
Download NeuCom Student
Kindly note that this version of NeuCom is for non-commercial use only. Please read through the License Agreement before you download and install NeuCom© Student.
Current version: v0.919 (released 21 Aug 2008)
- Dr. Qun Song
- Prof. Nik Kasabov
- Dougal Greer
- Peter Hwang
- Liang Goh
- Dr. Shaoning Pang