NotesFAQContact Us
Collection
Advanced
Search Tips
Peer reviewed Peer reviewed
Direct linkDirect link
ERIC Number: EJ913393
Record Type: Journal
Publication Date: 2010-Apr
Pages: 2
Abstractor: As Provided
Reference Count: 3
ISBN: N/A
ISSN: ISSN-0021-9584
A New Method for Studying the Periodic System Based on a Kohonen Neural Network
Chen, David Zhekai
Journal of Chemical Education, v87 n4 p433-434 Apr 2010
A new method for studying the periodic system is described based on the combination of a Kohonen neural network and a set of chemical and physical properties. The classification results are directly shown in a two-dimensional map and easy to interpret. This is one of the major advantages of this approach over other methods reported in the literature. The generated Kohonen map contains rich information on the relationships of chemical elements. This approach was not only able to group 54 elements into three clusters, metal cluster, semimetal cluster, and nonmetal cluster, but also offered detailed information on the (dis)similarity of the elements within each cluster. Furthermore, this study confirmed the existence of the singularity principle and the diagonal relationships. Finally, the trained Kohonen neural network was successfully applied to prediction of the properties of 5 test elements not included in the training set, a capability that may be employed to studying new chemical elements. This two-dimensional map-based method can be used as a complementary tool to the standard periodic table for both teaching and research. (Contains 1 figure.)
Division of Chemical Education, Inc and ACS Publications Division of the American Chemical Society. 1155 Sixteenth Street NW, Washington, DC 20036. Tel: 800-227-5558; Tel: 202-872-4600; e-mail: eic@jce.acs.org; Web site: http://pubs.acs.org/jchemeduc
Publication Type: Journal Articles; Reports - Descriptive
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A