Big data computational environment for tomography measurement data
Paper i proceeding, 2014
The authors propose a concept of computational environment for manipulation of big data sets originating from tomography-based measurement experiments. This work shows an example of utilizing the proposed tool in order to detect material plugs for pneumatic conveying measurement data. System is based on Hadoop distributed environment with Machout machine learning library. Paper presents results for a combination of horizontal and vertical flow experimental data coming from different experiments accumulated and processed in order to automatically detect plugs. Authors implemented supervised learning algorithms and naive Bayes classifier. Results show the possible way of using Big Data capabilities for both automatic data processing as well as for preparation of results for further analysis and interpretation by domain experts.
Measurement Data Processing
Machine Learning Alogrithms
Big Data Analysis