Page 18 - Big Data Analytics for Intelligent Healthcare Management

8 CHAPTER 1 BIO-INSPIRED ALGORITHMS FOR BIG DATA ANALYTICS

7 6
Number of papers 5 4 3 2 Evolutionary

Swarm-based

0 1 Ecological
2014 2015 2016 2017 2018
Year

FIG. 1.7
Time count of bio-inspired algorithms for big data analytics.

Type of analytics for bio-
inspired algorithms

Text analytics Audio analytics Video analytics Social media analytics Predictive analytics
Information LVCSR Server-based Content-based Heterogeneity
extraction analytics
architecture
Text Phonetic based Structure-based Noise
summarization system Edge-based analytics accumulation
architecture
Question Spurious
answering correlation
Sentimental Incidental
analysis endogeneity
FIG. 1.8
Type of analytics for bio-inspired algorithms.

The literature reported that there are five types of analytics for big data management using bio-
inspired algorithms: predictive analytics, social media analytics, video analytics, audio analytics,
and text analytics as shown in Fig. 1.8.
Text analytics is a method to perform text mining for an extraction of required data from the da-
tabase such as news, corporate documents, survey responses, online forums, blogs, emails, and social
network feeds. There are four methods for text analytics: (1) sentimental analysis, (2) question answer-
ing, (3) text summarization, and (4) information extraction. The information extraction technique ex-
tracts structured data from unstructured data, for example, an extraction of tablet name, type, and
expiry date from patient’s medical data. The text summarization method extracts a concise summary
of various documents related to a specific topic. The question answering method uses a natural lan-
guage processing to find answers to the questions. The sentiment analysis method examines the view-
point of people regarding events or products.
Audio analytics or speech analytics is a process of extraction of structured data from unstructured
audio data and examples of an audio analytics are healthcare or call center data. Audio analytics has