Logo
gamma_flow
Version
0.9.0
Short description
Guided Analysis of Multi-label spectra by Matrix Factorization for Lightweight Operational Workflows
Source catalogue
1
0
1
0
Description
In this project, a model is trained to analyze one-dimensional spectra, e.g. gamma spectra. It includes the preprocessing of spectra, training of a dimensionality reduction model and outlier analysis. As a result, (unlabelled) test spectra can be denoised and classified, including a classification uncertainty and a measure for the outlier probability. The model can be applied to single-label spectra as well as to multi-label spectra (linear combinations of the training spectra).
Features
- preprocessing of the spectra
- model training and testing for denoising and classification of spectra
- outlier analysis for detection of spectra unknown to the model
Detailed information
Release date
Supported languages
German,
English
Development status
Development
License
BSD 3-Clause License
Platforms
linux,
Microsoft Windows,
OSX
Maintenance type
[internal] Internally maintained by the repository owner
Software type
standalone/backend
Technical Contacts
UBA KI-Lab