Unscrambler Predictor, Unscrambler Models, Online Predictions of Samples

Allows reduction of manual calculations and testing procedures, therefore reducing the risk of error.

Can be customised inhouse or using CAMO's extensive expertise.

Transfer one single model across many sites or over different equipment in the same site.

Models developed in a central laboratory or technical group can be seamlessly transferred with no prior experience with The Unscrambler®. This leads to lower manufacturing costs and maximises ROI.

The Unscrambler Predictor is a CAMO initiative designed to make models generated off-line using The Unscrambler®, more widely accessible to third party instruments and control systems. The Unscrambler Predictor can be interfaced into an existing software platform to provide users all of the predictive and diagnostic tools found in The Unscrambler®, configured in a way that allows real time quality control.

Brochure

The Unscrambler Predictor "packages" an Unscrambler model into a DLL (32 bit only) protocol. Through these protocols, third party software, whether custom built (C, C++, Fortran, Visual Basic etc) or proprietary, can interface with the Unscrambler Predictor for the generation of real time prediction values and associated diagnostics.

Perform real time predictions from data obtained from laboratory or process equipment
DLL function calls(32 bit only)
Automatic data preprocessing, supports all pretreatments available in The Unscrambler®
Programmers reference manual available for in-house customisation
Supports PCR and PLS models built in The Unscrambler® 7.x + higher
Supports predictions from MLR models
Supports The Unscrambler®'s merged model file format for MLR, PCR and PLS-R. This format includes all model and pretreatment subfiles into a single file. This facilitates storage and re-location of The Unscrambler® model files for usage in Unscrambler Predictor
Outputs Hotelling T2 statistics, to help discover erroneous or outlying samples
Can be used together with Unscrambler Classifier(for online classification)
Runs under Windows 2000, XP, 2003 and Vista

Make predictions from raw data
Automatic preprocessing of new data
Works with several models simultaneously
Documentation and code examples follow the software

Providing quality measurements based on outputs of process spectrometers or other multichannel instruments
Detection of process deviations and faults before they become a problem
Pharmaceutical applications using Process Analytical Technology (PAT) and Quality by Design (QbD) initiatives
Reaction and process monitoring in the petrochemical, agriculture & food sectors and for environmental monitoring