Non-negative Tensor Factorization for Robust Exploratory Big-Data Analytics
- Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
Currently, large multidimensional datasets are being accumulated in almost every field. Data are: (1) collected by distributed sensor networks in real-time all over the globe, (2) produced by large-scale experimental measurements or engineering activities, (3) generated by high-performance simulations, and (4) gathered by electronic communications and socialnetwork activities, etc. Simultaneous analysis of these ultra-large heterogeneous multidimensional datasets is often critical for scientific discoveries, decision-making, emergency response, and national and global security. The importance of such analyses mandates the development of the next-generation of robust machine learning (ML) methods and tools for bigdata exploratory analysis.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- AC52-06NA25396
- OSTI ID:
- 1417803
- Report Number(s):
- LA-UR-18-20307
- Country of Publication:
- United States
- Language:
- English
Similar Records
Unsupervised machine learning based on non-negative tensor factorization for analyzing reactive-mixing
Discovering hidden geothermal signatures using non-negative matrix factorization with customized k-means clustering