 |
|
||||||
The Unscrambler Classifier is a CAMO initiative designed to make classification models generated in The Unscrambler®, more widely accessible to third party instruments and control systems. Classification of new samples is then possible in either an at-time or in-line mode of analysis and results can be displayed in tabular or graphical formats. |
The Unscrambler Classifier can be integrated into a client software platform by writing API function calls into it. Whether the software is developed in C/C++, Visual Basic, Fortran etc. functions are called in a DLL. New samples from the client software are fed into Unscrambler Classifier for reliable classification results. |
 |
- Perform real-time classification on raw data obtained from laboratories or process equipment
- SIMCA classification
- DLL function calls (32-bit) for easy usage
- Automatic data preprocessing, supports all pretreatments available in The UnscramblerŽ
- Programmers reference manual available for in-house customisation
- Supports The UnscramblerŽ's merged model file format for PCA. This format includes all model and pretreatment subfiles into a single file. This facilitates storage and re-location of The UnscramblerŽ model files for usage in Unscrambler Classifier
- Supports PCA models built in The UnscramblerŽ 7.x + higher
- Supports multiple models simultaneously
- Outputs Hotelling T2 statistics, to help facilitae the detection of outlying samples
- Can be used together with Unscrambler Predictor (for online prediction)
- Runs under Windows 2000, XP, 2003 and Vista
- Make classifications from raw data
- Works with The Unscrambler v 7.X + higher.
- Automatic preprocessing of new data
- Works with several sets of models simultaneously
- Documentation and code examples follow the software
- Provides quality measure used on the output of spectrometers or other process instruments
- Process surveillance, on-line quality assurance, automatically accepting or rejecting items
- On-line recognition and sorting of objects based on a combination of characteristics. for e.g. weight, size, shape etc
- Identify contaminated samples based on multiple measurements
- Classification of raw materials based on measurable characteristics
- Integration of spectral / chromatographic data of samples, which can be configured to classifysamples into groups like Good/Bad, Pure/ Adulterated etc
