Page 17 - Big Data Analytics for Intelligent Healthcare Management

1.3 BIO-INSPIRED ALGORITHMS FOR BIG DATA ANALYTICS: A TAXONOMY 7

Neeba and Koteeswaran [29] proposed a bacterial foraging optimization (BFO) algorithm to clas-
sify the informative and affective content from medical weblogs. MAYO clinic data is used as a med-
ical data source to evaluate the accuracy to retrieve the relevant information. Ahmad et al. [30]
proposed a BFO algorithm for network-traffic (BFON) to detect and prevent intrusions during the
transfer of big data. Further, it controls the intrusions using a resistance mechanism. Schmidt et al.
[31] proposed an artificial immune system (AIS) algorithm-based big data optimization technique
to manage and classify flow-based Internet traffic data. To improve the classification performance,
the AIS algorithm used Euclidian distance and the results demonstrate that this technique produces
more accurate results when compared to the Naı ¨ve Bayes classifier. George and Parthiban [32] pro-
posed the group search optimization (GSO) algorithm-based big data analytics technique using
FSO to perform data clustering for the high dimensional dataset. This technique replaces the worst
fitness values in every iteration of the GSO with the improved values from FSO to test the performance
of clustering data.

1.3.3 ECOLOGICAL ALGORITHMS
Pouya et al. [33] proposed the invasive weed optimization (IWO) algorithm-based big data optimiza-
tion technique to resolve the multiobjective portfolio optimization task. Further, the uniform design and
fuzzy normalization method are used to transform the multiobjective portfolio selection model into a
single-objective programming model. The IWO algorithm manages big data more quickly than PSO.
Pu et al. [34] proposed a hybrid biogeography-based optimization (BBO) algorithm for multilayer per-
ceptron training under the challenge of analysis and processing of big data. Experimental results show
that BBO is effective in providing training to multilayer perceptron and performs better in terms of
convergence when compared to the GA and PSO algorithm. Fong et al. [35] proposed the multispecies
optimizer (PS2O) algorithm-based approach for data stream mining big data to select features. An in-
cremental classification algorithm is used in the PS2O algorithm to classify the collected data streams
pertaining to big data, which enhanced the analytical accuracy within a reasonable processing time.
Fig. 1.6 shows the evolution of bio-inspired algorithms for big data analytics based on existing lit-
erature as discussed above.
Fig. 1.7 shows the number of papers published for each category of bio-inspired algorithm per year.
This helps to recognize the important types of bio-inspired algorithms [14–23, 35] [11–13, 25–29] that
were highlighted from 2014 to 2018.

2014 2015 2016 2017 2018
• CO [24] • FSO [14] • PCPSO [22] • SA [16] • ABC [13]
• GA [25] • PSO [19] • CSO [21] • SAFS [17] • ACO [26]
•BFON [35] • GSO [36] • SI [23] • PS2O [20] • BBO [10]
• IACO [27] • FSW [32]
•DE [28] • IWD [33]
• GP [29] • BFO [34]
• ES [30] •AIS [8]
• SFL [31] • FSOH [15]
•IWO [9]
FIG. 1.6
Evolution of bio-inspired algorithms for big data analytics.