Journal
COMPUTERS IN BIOLOGY AND MEDICINE
Volume 99, Issue -, Pages 161-172Publisher
PERGAMON-ELSEVIER SCIENCE LTD
DOI: 10.1016/j.compbiomed.2018.06.012
Keywords
Nanomaterials; Descriptors selection; Principal component analysis; Toxicity prediction
Categories
Funding
- Startup Foundation for Introducing Talent of NUIST [2017r103]
- R & D Program Development of User-friendly Nano-safety Prediction System - Ministry of Trade, Industry and Energy (MOTIE, Korea) [10043929]
Ask authors/readers for more resources
The present study is based on the application of a multivariate statistical analysis approach for the selection of optimal descriptors of nanomaterials with the objective of robust qualitative modeling of their toxicity. A novel data mining protocol has been developed for the selection of an optimal subset of descriptors of nanomaterials by using the well-known multivariate method principal component analysis (PCA). The selected subsets of descriptors were validated for qualitative modeling of the toxicity of nanomaterials in the PC space. The analysis and validation of the proposed schemes were based on five decisive nanomaterial toxicity data sets available in the published literature. Optimal descriptors were selected on the basis of the maximum loading criteria and using a threshold value of cumulative variance <= 90% on PC directions. A maximum inter-class separation(B) and the minimum intra-classes separation(A) were obtained for toxic vs. nontoxic nanomaterials in the PC space with the selected subsets of optimal descriptors compared to their other combinations for each of the datasets.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available