Advanced Analytics and Artificial Intelligence Applications


#### Table 5.

Summary with the best results of similarity.


Figures 9 and 10 show a comparison of results obtained from the algorithm with 100, 200, and 500 iterations of the instance of 1027 services. It can be seen that the best values of f(x) are produced with 500 iterations; however the result of the number of groups involved is more stable and better with 200 iterations.

This chapter describes a hybrid algorithm for Web services clustering; the approach is based on the ABC optimization algorithm combined with the K-means and Consensus. The clustering process starts from the extraction and processing of WSDL documents, then the calculation of semantic similarities between pairs of Web services operations. The semantic similarity measurements are based on WordNet, combining corpus-based and taxonomic approaches. As a result of the calculations, eight matrices are generated which are the input data set to the hybrid

Biologically inspired algorithms offer advantages over conventional clustering methods, such as the ability to find a near optimal solution by updating the candi-

The clustering algorithm designed is based on the behavior of the bees but was improved by incorporating K-means and Consensus so that the algorithm adjusts itself in each iteration. A series of experiments were conducted with the hybrid ABC algorithm, using optimal values for each adjustable parameter. The experiments were carried out with three collections with 50, 647, and 1027 Web services, and the algorithm was executed with variants of 100, 200, and 500 iterations. The hybrid

As future work, more combinations of semantic similarities, as well as incorporating more information of the description of the services and the data types incor-

date solutions iteratively and have self-organizing behavior.

Comparison of the number of groups in the best case for 1027 services.

Bio-Inspired Hybrid Algorithm for Web Services Clustering

DOI: http://dx.doi.org/10.5772/intechopen.85200

ABC algorithm has shown good results for Web services clustering.

8. Conclusions

Figure 10.

bio-inspired algorithm.

23

porated in the XML service definitions.

#### Table 6.

Summary with the best results of clusters.

Table 5 shows the results obtained with configurations: the best, worst, average, variance, and standard deviation of f(x) over the 20 executions of the algorithm (Table 6).

Figure 9. Comparison of similarity between the best solutions found with 1027 services. Bio-Inspired Hybrid Algorithm for Web Services Clustering DOI: http://dx.doi.org/10.5772/intechopen.85200

Figure 10. Comparison of the number of groups in the best case for 1027 services.

Figures 9 and 10 show a comparison of results obtained from the algorithm with 100, 200, and 500 iterations of the instance of 1027 services. It can be seen that the best values of f(x) are produced with 500 iterations; however the result of the number of groups involved is more stable and better with 200 iterations.

#### 8. Conclusions

Table 5 shows the results obtained with configurations: the best, worst, average, variance, and standard deviation of f(x) over the 20 executions of the algorithm

Comparison of similarity between the best solutions found with 1027 services.

Collection-iterations Best Worst Mean Variance Deviation –100 0.58936208 0.29274988 0.45901481 0.00593354 0.0770295 –200 0.54853117 0.44904535 0.50052478 0.00088301 0.02971549 –500 0.57163535 0.46554965 0.51820976 0.00090081 0.03001343 –100 0.54947584 0.2153186 0.44401451 0.01027133 0.10134758 –200 0.54778745 0.20257995 0.40923046 0.01554565 0.12468218 –500 0.54881541 0.22262872 0.41911288 0.01280278 0.11314939 –100 0.55531806 0.27378002 0.44488931 0.00847871 0.09207991 –200 0.56846422 0.28919387 0.47529505 0.00772078 0.08786795 –500 0.66112025 0.20302145 0.4674717 0.00977301 0.09885856

Advanced Analytics and Artificial Intelligence Applications

Collection-iterations Best Worst Mean Variance Deviation –100 2 10 4.31578947 6.22105263 2.49420381 –200 2 4 3 0.89210526 0.94451324 –500 2 6 3.94736842 2.36578947 1.53811231 –100 3 49 18.0526316 221.207895 14.8730594 –200 2 67 18.9473684 305.831579 17.488041 –500 2 73 23.2105263 426.368421 20.6486905 –100 2 110 31.2631579 1066.82895 32.6623475 –200 2 69 22.1052632 355.884211 18.8648936 –500 2 123 31 1179.21053 34.3396349

(Table 6).

Figure 9.

22

Table 6.

Table 5.

Summary with the best results of similarity.

Summary with the best results of clusters.

This chapter describes a hybrid algorithm for Web services clustering; the approach is based on the ABC optimization algorithm combined with the K-means and Consensus. The clustering process starts from the extraction and processing of WSDL documents, then the calculation of semantic similarities between pairs of Web services operations. The semantic similarity measurements are based on WordNet, combining corpus-based and taxonomic approaches. As a result of the calculations, eight matrices are generated which are the input data set to the hybrid bio-inspired algorithm.

Biologically inspired algorithms offer advantages over conventional clustering methods, such as the ability to find a near optimal solution by updating the candidate solutions iteratively and have self-organizing behavior.

The clustering algorithm designed is based on the behavior of the bees but was improved by incorporating K-means and Consensus so that the algorithm adjusts itself in each iteration. A series of experiments were conducted with the hybrid ABC algorithm, using optimal values for each adjustable parameter. The experiments were carried out with three collections with 50, 647, and 1027 Web services, and the algorithm was executed with variants of 100, 200, and 500 iterations. The hybrid ABC algorithm has shown good results for Web services clustering.

As future work, more combinations of semantic similarities, as well as incorporating more information of the description of the services and the data types incorporated in the XML service definitions.

Also the incorporation of other bio-inspired algorithms (swarm intelligence) for the classification and clustering of Web services, such as Particle Swarm Optimization (PSO), Bat Algorithm (BA), and Bird Swarm Algorithm (BSA).

References

[1] Liang Q, Li P, Hung PC, Wu X. Clustering web services for automatic categorization. In: SCC'09: IEEE International Conference on Services Computing. IEEE; 2009. pp. 380-387

DOI: http://dx.doi.org/10.5772/intechopen.85200

Bio-Inspired Hybrid Algorithm for Web Services Clustering

[9] Fellbaum C. WordNet: An Electronic Lexical Database. Cambridge, MA: MIT

disambiguation using machine readable dictionaries: How to tell a pine cone from an ice cream cone. In: Proceedings

[11] Banerjee S, Pedersen T. An adapted

[12] Wu Z, Palmer M. Verbs semantics and lexical selection. In: Proceedings of the 32nd Annual Meeting on Association

[13] Resnik P. Using information content to evaluate semantic similarity. In: Proceedings of the 14th International

[10] Lesk M. Automatic sense

Lesk algorithm for word sense disambiguation using WordNet. In: Computational Linguistics and Intelligent Text Processing. Berlin Heidelberg: Springer; 2002. pp. 136-145

for Computational Linguistics. Association for Computational Linguistics; 1994. pp. 133-138

Joint Conference on Artificial Intelligence. Montreal; 1995.

[14] Jiang JJ, Conrath DW. Semantic similarity based on corpus statistics and lexical taxonomy. In: Proceedings of International Conference on Research in Computational Linguistics, Taiwan. 1997

[15] Lin D. An information-theoretic definition of similarity. In: Proceedings of the 15th International Conf. on Machine Learning. San Francisco, CA: Morgan Kaufmann; 1998. pp. 296-304

[16] Hirst G, St-Onge D. Lexical Chains as Representations of Context for the

Combining local context and WordNet

Detection and Correction of Malapropisms; 1998. pp. 305-332

[17] Leacock C, Chodorow M.

pp. 448-453

of SIGDOC '86. 1986

Press; 1998

[2] Platzer C, Rosenberg F, Dustdar S.

multidimensional angles as proximity measures. ACM Transactions on Internet Technology. 2009;9(3):11

Computation: Practice and Experience.

[5] Wu J, Chen L, Zheng Z, Lyu MR, Wu Z. Clustering web services to facilitate service discovery. Knowledge and Information Systems. 2014;38(1):

[6] Prakash J, Singh PK. Evolutionary and swarm intelligence methods for partitional hard clustering. In: 2014 International Conference on

Information Technology (ICIT). IEEE;

[7] Sahoo G. A two-step artificial bee colony algorithm for clustering. Neural Computing and Applications. 2017;

[8] Kotekar S, Kamath SS. Enhancing web service discovery using metaheuristic CSO and PCA based clustering. In: Progress in Intelligent Computing Techniques: Theory, Practice, and Applications. Singapore: Springer; 2018.

[4] Du YY, Zhang YJ, Zhang XL. A semantic approach of service clustering and web service discovery. Information Technology Journal. 2013;12(5):967-974

Web service clustering using

[3] Pop CB, Chifu VR, Salomie I, Dinsoreanu M, David T, Acretoaie V, et al. Biologically-inspired clustering of semantic web services. Birds or ants intelligence? Concurrency and

2012;24(6):619-633

207-229

2014. pp. 264-269

28(3):537-551

pp. 393-403

25
