Page 16 - Big Data Analytics for Intelligent Healthcare Management

6 CHAPTER 1 BIO-INSPIRED ALGORITHMS FOR BIG DATA ANALYTICS

1.3.2 SWARM-BASED ALGORITHMS
Ilango et al. [9] proposed an artificial bee colony (ABC) algorithm-based clustering technique for man-
agement of big data, which identifies the best cluster and performs the optimization for different dataset
sizes. The ABC algorithm approach minimizes the execution time and improves the accuracy. A
MapReduce-based Hadoop environment is used for implementation and results demonstrate that the
ABC algorithm delivers a more effective outcome than the differential evolution and particle swarm
optimization (PSO) in terms of execution time. Raj and Babu [18] proposed a firefly swarm optimi-
zation (FSO) algorithm for big data analytics for establishing novel connections in social networks to
calculate the possibility of sustaining a social network. In this technique, a mathematical model is in-
troduced to test the stability of the social network and this reduces the cost of big data management.
Wang et al. [19] proposed an FSO algorithm-based hybrid (FSOH) approach for big data optimization
to focus on six multiobjective problems. It reduces execution costs but it has high computational
time complexity.
Wang et al. [20] proposed a PSO algorithm-based big data optimization approach to improve online
dictionary learning and introduced a dictionary-learning model using the atom-updating stage. The
PSO algorithm reduces the heavy computational burdens and improves the accuracy. Hossain et al.
[21] proposed a parallel clustered PSO algorithm (PCPSO)-based approach for big data-driven service
composition. The PCPSO algorithm handles huge amounts of heterogeneous data and process data
using parallel processing with MapReduce in the Hadoop platform. Lin et al. [22] proposed a cat swarm
optimization (CSO) algorithm-based approach for big data classification to choose characteristics dur-
ing classification of text for big data analytics. The CSO algorithm uses the term frequency-inverse
document occurrence to improve accuracy of feature selection.
Cheng et al. [23] proposed a swarm intelligence (SI) algorithm-based big data analytics approach
for the economic load dispatch problem and the SI algorithm handles the high dimensional data, which
improves the accuracy of the data processing. Banerjee and Badr [24] proposed the ant colony opti-
mization (ACO) algorithm-based approach for mobile big data using rough set. The ACO algorithm
helps to select an optimal feature for resolved decisions, which aids in effectively managing big data
from social networks (tweets and posts). Pan [25] proposed the improved ACO algorithm (IACO)-
based big data analytical approach for management of medical data such as patient data, operation data
etc., which helps doctors retrieve the required data quickly.
Hu et al. [26] proposed a shuffled frog leaping (SFL) approach to perform the selection of the fea-
ture for improved high-dimensional biomedical data. For improved high-dimensional biomedical data,
the SFL algorithm maximizes the predictive accuracy by exploring the space of probable subsets to
obtain the group of characteristics and reduce irrelevant features. Manikandan and Kalpana [27] pro-
posed a fish swarm optimization (FSW) algorithm for feature selection in big data. The FSO algorithm
reduces the combinatorial problems by employing the fish swarming behavior and this is effective for
diverse applications. Social interactions among big data have been designed using the movement of
fish in their search for food. This algorithm provides effective output in terms of fault tolerance
and data accuracy. Elsherbiny et al. [28] proposed the intelligent water drops (IWD) algorithm for
workflow scheduling to effectively manage big data. The workflows simulation toolkit is used to test
the effectiveness of the IWD-based approach and results show that the IWD-based approach is per-
formed effectively in terms of cost and makespan when compared to the FCFS, Round Robin, and
PSO algorithm.