You are hereData Analysis
Data Analysis
Data analysis for chemical sensors requires a good knowledge about the properties of the gas sensors, the requirements of the specific application and the capabilities of the various algorithms.
Within our various software products we offer a broad range of algorithms that are especially suitable for multi sensor systems.
Feature extraction
Feature extraction is used to extract a limited amount of values from a sensor response curve. Typical features are the maximum signal, baseline corrected signals, derivative signals or relative signals, but may also be the time needed to reach the maximum or other dynamic properties.
Filters
Filters are typically used to remove noise from measurement data. Filters may also be used to remove unwanted frequency domains from a sensor signal, adopt to changing baseline or convert data into other domains (e.g. time/frequency domains).
Multivariate data analysis
Multivariate data analysis algorithms are used to visualize data, classify or predict based on reference data. Typical algorithms that we offer are:
- Principal Component Analysis (PCA)
- Linear Discriminant Analysis (LDA)
- Principal Component Regression (PCR)
- Partial Least Squares (PLS)
- Backpropagation Neural Networks
A complete set of these algorithms is available via our desktop software packages. Embedding these algorithms into our instruments is done for specific applications in close colaboration with our customers.