ClasSOMfier: A neural network for cluster analysis and detection of lattice defects
Project description
ClasSOMfier: A neural network for cluster analysis and detection of lattice defects
Class that classifies atoms according to their environment. Unsupervised training using a 1-dimensional Self Organizing Map (SOM) in Fortran.
Created by Javier F. Troncoso, October 2020. Contact: javierfdeztroncoso@gmail.com
USE:
The network and its parameters can be initialized using the following commans:
>>nn=ClasSOMfier(6.43718,2,"dump1000.file")
Only 3 parameters are necessary: characteristic length, number of clusters and input file.
The format of the input file is that provided by the dump command in LAMMPS:
#compute peratom all pe/atom
#dump dumpid2 all custom 1000 dump*.file id mass x y z c_peratom
The first command calculates and stores the potential energy per atom.
The network is trained using the following command:
>>nn.execute()
The final condigurations are written in ./data (default value) and can be easily read by Ovito.
The final configuration can be postprocessed so that it can be used again to find subcategories
inside a specific category:
>>nn.postprocess_output()
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
classomfier-0.1.tar.gz
(4.3 kB
view hashes)
Built Distribution
classomfier-0.1-py3-none-any.whl
(10.3 kB
view hashes)
Close
Hashes for classomfier-0.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 171be2175aeb09215d419fb04aacd83fc87bba861d8c3932ba97a862c02526d0 |
|
MD5 | 26def5e376e79c605ae9d575d97b5893 |
|
BLAKE2b-256 | 49944185ebef0bf5e8698b584a710dc1c34f74abe8f238f8df5fdbecd87eaf7e |