The effectiveness of hybridization mechanism of IHDE-EDA is first discussed, and then simulation and comparison based on three benchmark problems demonstrate the efficiency, accuracy and robustness of IHDE-EDA.
Simulations and comparisons based on twenty benchmark problems and four real-world engineering designs demonstrate the effectiveness, efficiency and stability of NBA compared with the basic BA and some well-known algorithms, and suggest that to improve algorithm based on biological basis should be very efficient.
Based on three benchmark test collections Cranfield, Medline and CACM, experiments have been carried out to compare our method with three well-known query optimization methods on relevance feedback: the traditional Ide Dec-hi method, the Horng and Yeh's GA-based method and the López-Pujalte et al.'s GA-based method.
Figure 4 shows the resulting AUC scores and the number of inferred features by L1LOG and L2LOG based on nine benchmark datasets with different clustering thresholds (see the Method section for more details).
Finally, based on four benchmark datasets, we demonstrated the effectiveness of our approach, by comparing with NBI, BLM, and BLM's extensions in cross validation and predicting potential interactions in unapproved DTPs under checking in existing databases.
For example, the pharmacogenomic approach worked well for the TestDrug-TrainProtein pairs when the chemical similarity threshold is 0.2, and for the TestDrug-TestProtein pairs when the chemical similarity threshold lies in the range 0.2-0.5 0.2-0.5 5 shows the resulting AUC scores and the number of inferred features for L1LOG-C and L2LOG-C based on nine benchmark datasets.
In the recently published Innocenti Report Card no. 8 by UNICEF (2008), Hungary is 8th among 25 OECD countries in terms of provision of early childhood care, based on ten benchmarks related to four broad categories: policy framework, access to services and their quality, and supporting context.
There is a 2% absolute increase (5% relative improvement) in alignment accuracy over SP4 based on two independent benchmarks.
The study was based on two well-known benchmark software product lines: MobileMedia (Young 2005) and Berkeley DB (Kästner et al. 2008).
