Biodiv Sci ›› 2002, Vol. 10 ›› Issue (3): 345-350.  DOI: 10.17520/biods.2002048

• 论文 • Previous Articles    

Functional Link Artificial Neural Networks and agri-biodiversity analysis

ZHANG Wen-Jun,QI Yan-Hong,Schoenly K G   

  1. (中山大学生物防治国家重点实验室,生命科学学院,计算机科学系,  广州 510275)
    (Department of Biological Sciences , California State University , Stanislaus , CA 95382 , USA)
  • Received:2001-12-11 Revised:2002-04-25 Online:2002-08-20 Published:2002-08-20
  • Contact: ZHANG WenJun

Abstract: A computational software using Functional Link Artificial Neural Network (FLANN) was developed. This software includes seven Java classes and one HTML file. It is an Internet online computational software which can be run on various operational systems and web browsers, and can read various databases. Two data sets of rice insect biodiversity were lumped and clustered employing the biodiversity software LUMP and hierarchical clustering, which resulted in two sets of training samples each with 60 samples and, 21 and 20 functional groups, respectively. Ten samples to be predicted were prepared for each set of data. FLANN was used to make pattern classifications. The simulated and predicted results with FLANN software gave good fitness to the investigated results. The development of FLANN software also provides a new tool for the development of Internet intelligent analytical system for biodiversity studies.