Information Technology Reference
In-Depth Information
(
Continued
) (b) Conventional and nonconventional
formulation classes. Compound descriptors used
are: number of hydrogen bond donors and acceptors
(HBD and HBA, respectively); % polar surface area
(%PSA); logarithm of dose and dose number (logD
and logD
0
, respectively); molecular surface area
(reprinted from Branchu et al., 2007; with
permission from Elsevier)
Figure 5.6
Example 2
(Zucker et al., 2009)
A decision tree was developed to aid selection of candidate
molecules for liposome loading and to optimize loading conditions.
The tree is constructed on the basis of physicochemical principles
and connects the drug properties, loading conditions, and loading
effi ciency. The J48 tree was built on the basis of the C4.5 algorithm,
and is a J48 open source Java implementation of the C4.5 algorithm
in the weka data mining tool. Weka is available on the University of
Waikato website (
http://www.cs.waikato.ac.nz/
∼
ml/weka/
).
Interpretation of the decision tree can be done in the following
way: the most important loading condition that affects loading
effi ciency in the decision tree (Figure 5.7) is the initial drug/lipid
Search WWH ::
Custom Search