Information Processing and Management 59 (2022) 102762

Contents lists available at ScienceDirect

Information Processing and Management

journal homepage: [Link]/locate/infoproman

Big data-assisted social media analytics for business model for

business decision making system competitive analysis
Honglei Zhang a, Zhenbo Zang b, *, Hongjun Zhu c, M. Irfan Uddin d, M. Asim Amin e
a
Research Center of Wuling Mountain Area Characteristic Resources, Development and Utilization, College of Finance and Economics, Yangtze
Normal University, Chongqing 408100, China
b
School of Economics and Management, Chongqing Metropolitan College of Science and Technology, Chongqing 402167, China
c
Graduate Admissions Office, Civil Aviation Flight University of China, Guanghan 618307, China
d
Institute of Computing, Kohat university of science and technology, Kohat, Pakistan
e
Tsinghua-Berkeley Shenzhen Institute, China

A R T I C L E I N F O A B S T R A C T

Keywords: Business is based on manufacturing, purchasing, selling a product, and earning or making profits.
Social media Social media analytics collect and analyze data from various social networks such as Facebook,
Data analysis Instagram, and Twitter. Social media data analysis can help companies identify consumer desires
Companies
and preferences, improve customer service and market analytics on social networks, and smarter
Big data
Customer
product development and marketing investments. The business decision-making process is a step-
Decision-making by-step process that enables employees to resolve challenges by weighing evidence, evaluating
possible solutions, and selecting a route. In this paper, Big Data-assisted Social Media Analytics
for Business (BD-SMAB) Model increases awareness and affects decision-makers in marketing
strategies. Companies can use big data analytics in many ways to enhance management. It can
evaluate its competitors in real-time and change prices, make deals better than its competitors’
sales, analyze competitors’ unfavorable feedback and see if they can outperform that competitor.
The proposed method examines social media analysis impacts on different areas such as real
estate, organizations, and beauty trade fairs. This diversity of these companies shows the effects of
social media and how positive decisions can be developed. Take better marketing decisions and
develop a strategic approach. As a result, the BD-SMAB method enhance customer satisfaction
and experience and develop brand awareness.

1. Introduction to social media and business decision model

Organizations face many obstacles in numerous business fields, such as decreasing finances, rising expenses, increasing complexity,
and increased service needs (Gao, Wang, & Shen, 2020). Many firms take up knowledge management (KM) for better decision-making
to overcome this situation (Zong, Yuan, Montenegro-Marin & Kadry, 2021). Global corners in a company around the globe differ from
an international firm that sells products worldwide; a marketer’s knowledge has facilities in its own country. It led to some KM actions
and techniques to improve company performance, increase creativity, and achieve organizational goals. Knowledge management is
essential if companies develop and preserve a comparative edge over time (Gheisari et al., 2021b; Zhao, Yu, Shakeel &

* Corresponding author.
E-mail addresses: zhanghonglei@[Link] (H. Zhang), zhanghl_rambo@[Link] (Z. Zang), hongjunzhu664@[Link] (H. Zhu), irfanuddin@
[Link] (M.I. Uddin), asim20@[Link] (M.A. Amin).

[Link]
Received 15 June 2021; Received in revised form 4 September 2021; Accepted 8 September 2021
Available online 27 September 2021
0306-4573/© 2021 Elsevier Ltd. All rights reserved.
H. Zhang et al. Information Processing and Management 59 (2022) 102762

Montenegro-Marin, 2021). Now that customers are ready to pay much to these data, which may help them target specific market
segments, they have become big.
Managing an organization’s social media information for business objectives is a new problem nowadays in business-focused
knowledge management. Social media analysis of information for consumer web browsing, online shopping, customer feedback,
and social-network marketing research allows organizations to gain timely and comprehensive views of platforms. Organizations
compete to connect Big Data initiatives with organizational learning infrastructure to respond to these new Big Data and social media
developments (Kumar et al., 2020; Pham, Nguyen, Nguyen, Pham & Nguyen, 2020). Big data analyses will better enable marketers to
understand their online communities and predict their behaviors to provide customized services and solve any problems quickly. These
activities are generally aimed at capturing, saving, retrieving, and managing all feasible information from the social networking
material provided by users relevant to their organizations and rivals (Golpayegani, Esmaeili, Mardani, & Mutallebi, 2019). Database
management network links tasks, decisions, information, and data supply based on standardized models and notes and domain-specific
data models. CEOs may make better management decisions based on the insights generated from longitudinal social media data (Song
et al., 2019). The BD-SMAB model essentially comprises four steps of the suggested decision-making procedure to enable information
requirements to be derived and determine the impact on incoming observations’ relevant tasks and decisions.
The knowledge gained an understanding of how clients communicate about their services on social networking sites, such as
Facebook or Twitter (Al-Turjman & Salama, 2020). The intelligence stage begins with six phases: demand, planning and management,
collection, processing and use, analysis and production, and dissemination. Recent research by Bain & Company showed that major
corporations with sophisticated analytical capacities and big data technology beat their competition. These choices are defined with
active collaboration, and the cycle in nature is circular. The movement involves the essential facts, knowledge to be derived, and prior
decisions essential to make choices. They noted that big data analysis users had taken smart judgments five times quicker than the
competition (Chandrasekaran, Nguyen & Hemanth, 2021). Because of the business information derived by big data analysis, the
company is significantly more likely to be in the highest four in the finance industry. The selection of forms and icons for graphical
elements identified in this specification is a key element of DMN. While sponsorship is plentiful in media studies and big data studies,
practical assistance for big data integration, social media, and organizational learning is sparse (Balaanand et al., 2019). The method of
feeling analysis is developed to analyze organizations and complaints by grouping and scattered comments for basic sentences. Thus,
the objective of this study is to check the usefulness of the development and implementation of an integrative management information
system. With the rapid rise of reviewing and social media websites, service knowledge based on user reviews is accessible (Gheisari
et al., 2021a). Practical examination of these evaluations can give insight into the products/services that customers like/dislike. This
stage has been broken into two sub-tasks: decision comprehension and modeling and data collection and processing to create a
standard visual language recognized and understood by all decision-makers.
Unstructured information on the review is significant, speed and diverse, and challenging to deal with for business information and
decision-making using standard methodologies, management instruments, and processes (Nie et al., 2020).In particular, the decision
of comprehension for dynamic characteristics of a framework for business users who wish to understand the model of the process. DMN
features represent the characteristics of the user who the organization of workshop models influences. The current methodologies take
the whole corpus into account for evaluation without examining the credibility of the examinations. Different difficulties have been
identified by analytics in the prior evaluation (Ngan et al., 2019). Because DMN allows you to turn business rules into knowledge assets
where processes must be separated from corporate rules. This occurs when the sequence of activities to produce the result differs from
the rules in the process. This post utilized customer feedback data to assess client satisfaction/concerns based on customer evaluations.
The stage that aims to utilize decision-making involves extensive, limited making, decision, and routine decision-making processes.
Knowledge management is a means of collecting, distributing and using information from a company to make it more available to
everyone. When properly implemented, a knowledge management system is the only source of truth in a company, making it a source
for accurate decision-making.
The social data of the customer is pre-processed and divided into phrases. The phrase score of the Affin lexicon was calculated
(Yang et al., 2016). Naturally, as companies use Facebook, Linked In, and Twitter progressively as marketers in this day and age, the
competitions can retrieve some intriguing details – and perhaps even the own company – simply by tuning them in to surpass them.
The goal of communication and cooperation involves various factors that influence public policies, such as public opinion, eco­
nomic conditions, new science, technological changes, interests groups, NGOs, corporate lobbying, and political activity. The phrases
are screened out without information or neutrality, and the Ensembles Classification based on machine learning using Hybrid char­
acteristics is utilized to identify phrases as praise/complaint. The construction of this process relies on an ongoing cycle of planning,
analysis, implementation, testing, and evaluation at the most basic level. The iterative process begins with initial planning and global
demands, unlike a Waterfall approach. Social media analytics uses several methodologies designed to generate certain metrics from
social media data. Researchers discover that sentiment analysis, social networking analytics, and statistical methodologies are most
commonly employed, based on a survey of current commercial instruments.
The sentiment analysis methodology is created to analyze praises and complaints by grouping and scattered parsing to obtain basic
sentences (Voramontri & Klieb, 2019). For the clustering of appreciations/complaints, the nonnegative vector factorization approach
is utilized. Top-ranking compliments and comments are shown with their word patterns to acquire vital insights and decisions.
The contributions of this research are given below:

• Intelligent phase, detection phase, and design phase to analyze the social media message.
• Social media hotness / social media index score is proposed to find the brand value in the social media.
• The proposed model is analyzed with a case study.

2
H. Zhang et al. Information Processing and Management 59 (2022) 102762

The rest of the research is as follows. Section 2 deals with the background of social media and the business decision model. The
proposed Big data-assisted social media analytics for business (BD-SMAB) model is designed and implemented in Section 3. The
software analysis and performance evaluation are discussed in Section 4. The conclusion and future scope are illustrated in Section 5.

2. Background to the social media and business decision model

Several systems have already been deployed to analyze fine grain or analyze crucial sentences. These algorithms take the complete
corpus of analysis into account and grammatically inform a phrase (Kovacova, Kliestik, Pera, Grecu & Grecu, 2019). It is supposed to
contain the word and adjectives in a subjective statement. Present-day customers assess, choose and exchange their knowledge with
service/product internet has shifted from material provided by users. Traditionally, most of the study is based on customers’ quan­
titative assessments on Internet portals. To gauge customer happiness, Advanced Analytics can collect helpful information from useful
evaluations (Alhakimi & Alwadhan, 2021). Methods for Big data analysis may be utilized for knowledge management, company
analytics, customer experience, and decision-making in virtual environments.
Noureddine et al. have investigated the characteristics of commendation and complaints, where commendation is a sub-set of
favorable only evaluations (Ma, Au & Ren, 2020). The Louise Sentence is lengthier than the Average Statement and has more Verb,
Expressions, and Nouns. The phrase includes fewer adjectives and pronouns, simpler past and prepositions, instead of the complaint
(Shahrani & Ghandour, 2018). Existing sentiment analytical algorithms can be enhanced solely by analyzing language from large
customer evaluations of complaints.
Hossain et al. have devised a summary method of opinion based on language characteristics of review phrases (Noureddine &
ZeinEddine, 2018). They have merely analyzed the overall assessment and retrieved opinions without studying the positive or negative
aspects of the viewpoint phrases (Hansen, Saridakis & Benson, 2018). The review ranking method Tafesse et al. have designed is based
on thoughtfulness (Hossain, Dwivedi, Chan, Standing & Olanrewaju, 2018). It identifies high-quality assessments based on hybrid
aspects such as substances, adjectives, verbs, hard phrases, incorrect phrases, volatility, and random-forest rankings. Low-quality
comments are disregarded, and reputable reviews are shown exclusively (Chang, Ku & Chen, 2019). Research shows that sentiment

Fig. 1. The architecture of the proposed Big data-assisted social media analytics for business (BD-SMAB) model.

3
H. Zhang et al. Information Processing and Management 59 (2022) 102762

analysis is used to analyze consumer evaluations of the facility. Ghani et al. have researched the characteristics using text analysis and
its effect on people’s impressions of the town of Los Angeles (Tafesse & Wien, 2018).
A map was constructed and examined, depending on the attributes of the facility. Oksana et al. examined quality criteria and
models of various services (Bhandari & Bansal, 2019). They created a quality assessment methodology based on numerous charac­
teristics. Bezzera has evaluated a six-factor terminal measuring service quality methodology Check-in, safety, accessibility, atmo­
sphere, basic and comfortable facilities (Ghani, Hamid, Hashem & Ahmed, 2019). Information from the Brazil company was acquired
on-site. Their performance cannot be generalized to all locations, given their personal and contextual perception of customer
satisfaction.
Genuine and meaningful feedback from consumers is created by users, freely open and reachable wherever and whenever. There is
no standardized, organized approach for obtaining and analyzing user-generated content to develop management decision knowledge
(Oksana, 2020). Therefore, this methodology finds consumer satisfaction findings based on consumer shared compliments and
grievances using Big Information from unorganized textual user evaluations. The most major distinction between previous research
and the research wants to use the standards with big data to facilitate interaction among decision-makers and Big Data analytics teams
during decisions throughout the decision-making procedure (Liang & Liu, 2018). The standard provides an understandable repre­
sentation of human choice. It gives a consistent notation that all enterprise customers and the research team may readily grasp.
The big data-assisted social media analytics for business (BD-SMAB) model has been built intelligently to describe decisions and
needs. The decision-modeling utilizing the format consists of two layers, the judgment demands diagrams and the decision-making
logic. Decisions include several aspects and their interconnections. These defined the judgment and how it varies depending on
other decision making, policy proposals or regulations (acquisition of information), business skills (understanding model), and in­
formation on elements. At the same time, judgment logic sets out the logic that allows the validation/automating of the decision-
making procedures to start making independent choices, including business policies, decision graphs, or executable numerical
simulations.

3. Proposed Big data-assisted social media analytics for business (BD-SMAB) model

Thanks to the Internet of Things, businesses can now collect data better to understand their consumers’ buying and behavioral
trends. Technological advancements assist scientists in dealing with global concerns and give marketers the knowledge they need to
make informed decisions. Social media, the rise of social media marketing, and the availability of big data on social platforms have
assisted us in better understanding how smart technology may affect our lives shortly. Marketers can utilize big data to discover social
media trends and acquire insights, which can be utilized to make engagement decisions such as which people to contact, which set of
users should get marketing emails, etc.
Fig. 1 shows the proposed Big data-assisted social media analytics architecture for business (BD-SMAB) model. The architecture,
database management network (DMN) standards and integrative model were developed based on several earlier research works. BD-
SMAB Model essentially comprises four steps of the suggested decision-making procedure: information stage, conceptual design,
development, and implementation stage.
The methodology begins with the intelligence stage when the surroundings are surveyed for the choice to be taken. This choice and
its needs are defined at this phase. These criteria comprise the essential facts, the knowledge to be derived, and the results of prior
decisions essential to make this choice. It suggests designing the choice using the DMN standards to detect and comprehend all these
needs. This step also involves the collection and preparation of the necessary information. This stage has been broken into two sub-
tasks: decision comprehension and modeling and data collection and processing.

• Decision of comprehension and modeling: the method is described and comprehended, and then the framework is reflected by the
standards of DMN. It recommends using the DMN standards in this stage to shape human choices. Individual decisions may be
divided into a system of parts that describe judgments and their needs. This stage aims to utilize decision-making as an answer to,
simply, and unambiguously express decision-making. The goal is also to enhance communication and cooperation among poli­
cymakers and the analytical group.

Hence, it may use everyday language instead of decision logic to construct. If this choice is to be automated, it employs a planning
stage to define. The construction of the model is carried out through the iterative application of four stages, the suggested decision
identification, decision description, decision-making criteria, and design break-down and refining. The source in the previous module
identifies the inputs to be examined for collecting the necessary information. This stage is required to identify the various quantitative
and qualitative methods that provide this data.

• Data Gathering and Planning: this substage is comparable to identifying big data, the acquisition/store of data, the organization of
processes specified during the analysis. In the preceding sub-phase, the inputs to be studied to collect the necessary data are
determined by the source. This process phase is necessary to identify the different primary and secondary sources that give these
data. These data must then be collected and saved from these providers. This phase of the process is essential to recognize the
different sources of descriptive and inferential statistics. These data must be collected from these services and recovered from them.

The data can be available in many varieties: images, words, multimedia, audio files, etc. Big data technology offers several options
supporting these phases, such as Hadoop, Chukwa, Afrojack, and Kino for data acquisition from a separate supplier to be placed in

4
H. Zhang et al. Information Processing and Management 59 (2022) 102762

HDFS/NoSQL repositories as Hbase, MongoDB, Jdbc, Ruby, etc. The information may then be analyzed, searched, and shown to obtain
a basic overview and analysis using several tools, includingSparkSQL, Hadoop, for information querying SparkR for statistical ap­
plications, Dashboard for display, etc. The data can be available in many forms, where a variety of Big Data refers to all structured and
unstructured data produced by people or by machines. The ability to classify incoming data into different categories is important.
The second stage is the planning stage when various action paths are devised, created, and examined to deal with decision-making.
This stage can alternatively be divided into two sub-stages: building models and generating solutions. The stages can be divided based
on various pathways of action that are developed and considered to address the decision-making issue.

• Build prototype: when models have been developed to retrieve the necessary knowledge for that choice for Big Data Technology,
that knowledge is currently recognized and reflected in the framework of decision needs. Two stages in the construction of a
model are model design and model development. A prototype is used to developing the product’s preliminary model. Using
prototypes involves testing an idea’s usefulness, convenience, or viability. A methodology or a short selection of possible ap­
proaches to developing the model are selected in the model preparation. The information is examined to determine character­
istics appropriate for the system. Big data helps enterprises gain precious insights. Companies use big data to refine their
marketing campaigns and techniques further. The companies use this to train machines, predictive models, and other advanced
analytical applications in machine-learning projects. The strategies to be used are selected based on the design aim and the
information type provided. The stages can be divided according to various action paths and discussed to deal with the stated
problem involving interaction design and usability.

Afterward, the databases are produced, constructed, and carried out in a building design for evaluation, learning, and
manufacturing. The model construction is developing a similarity measure that best describes the relation of dependent to independent
variables. The major problems are how the relationship is formed and what independent variables are selected. Several analysis ap­
proaches might be utilized in this stage to develop simulation solutions, such as categorization of data mining methods, clustering,
modeling, classification methods, machine training methods, text classification, data visualization, and recommender systems. Many
big data technologies such as Spark Mahout and so on may be utilized in this stage. A system is a neat grouping of interdependent
components linked to a particular objective according to a plan. The components have to have interrelationships and interdependence.
The aims of the whole organization are more important than the goals of its subsystems.

• Produce options: decision-maker depend on new information, expertise and practice to provide options available in this sub-
phase. In addition, criteria are set for the evaluation and each choice. Several approaches for analysis could be used in this
phase to create solutions for simulation, including the classification of statistical approaches, grouping, and simulations. The next
phase is the selection step. These options are analyzed to determine each effect on the company after establishing alternatives and
assessment criteria. The standard offers to assess alternatives by introducing a server to forecast the most appropriate solutions
amongst these provided options. It uses memory caching and optimized query performance for quick queries of any size. Spark is
a quick and general engine for the processing of large-scale data.

The visualization process may evaluate alternatives via monitoring, displays, simulations, modeling of answers, and other ap­
proaches. Based on the assessment results, decision-makers pick one option or several options to fix the issues. Therefore, the last level
of the decision-making procedure is the step of implementing the preferred choices.
The linear representation of the cost is denoted in Eq. (1)
∑
K
Δxk (t)
ΔC(t + 1) = wk (1)
k=1
k

It can calculate its search terms, xk (t), at period t, officially for each of K phrases. This quantity is suggested to affect future share
prices, C(t +1), at time t + 1. It contrasts the market value changes, ΔC(t + 1) at a period t + 1 with changes in phrase-based intensity
at t period Δxk (t). wk is denoted the weight of every phrase. Weights are calculated with the optimal phrases with the highest non-zero
costs for each phrase. The weight is calculated using Eq. (2)
( ( )2 )
∑ ∑K
Δxk (t) ∑⃒⃒wk ⃒⃒
a = (a1 , a2 , ⋯, aK ) = argmax ΔC(t) − wk +λ ⃒ ⃒ (2)
t k=1
k k
k

Based on the reliability of final forecasts, λset from practical assessments. k is the number of brands, and C(t) is the price of the
brand product or service. The weight of the final big data is denoted as a and which is given as a1 ,a2 ,⋯,aK .Since Search engine numbers
are only provided occasionally, the broader trend of public inquiries can be masked. The zscore(5) of Facebook’s search traffic between
each fortnight as the average value the daily of that fortnight overcomes this issue. It utilizes a score(25) of each daily net asset value to
measure the price levels that fluctuate daily. A zscore is set by Eq. (3)
X − Mx
zscore(n) = ∑ (3)
n∗

5
H. Zhang et al. Information Processing and Management 59 (2022) 102762

∑
If X is the 1-day variation, Mx is denoted the average return of 1-day variations of trailing n-day and is the point distinction of
these moving 1-day variations of n-day. It is aligned with the daily pricing data for Search Engine Analytics. It also guarantees that
deviations are evaluated for approximately the same period: around one month. Creating capabilities that correlate Facebook’s pat­
terns with banking sector values makes it possible to understand how variations in search volumes mirror market fluctuations.
Burst characteristics identification: It wants to see forecasts for national stock exchanges. It must know which state was responsible
for creating a tweet. Although tweets may contain altitudes, and lengths sent by individuals, most tweets do not include this infor­
mation. To avoid this, it used the geo-enriched tool mentioned above to apply another geo-enriching technique to the database. Once
tweets are available for a nation of relevance, it can cover the period and compute each nation’s burst rating.
The frequency-inverse occurrence of the material method is used to calculate each explosion rating word. When t exceeds a limit,
the word |zscore(30)| was used to indicating bursting characteristics. Preparing for a Blast, it gets a sequence of Bursting vectors
represented by B = {b0 ,b1 ,⋯,bl }.Following computation of Bursting numbers for each term. If the C coefficients between the two edges
are high enough, it can build a graphing model, G = (D, F, P), to connect nodes to an edge in F. Where feature denotes the complex
conjugate of D and P. The number of interconnections is calculated using Eq. (4)
∑
∞
f [p] ∗ g[q] = f ∗ [p]g[q + p] (4)
m=− ∞

f[p] and g[q] are denoted the two functions. f ∗ [p] is the complex conjugate of the function f[p]. For the requirements, vectors with a
cross-relationship rating of 0.1 are considered closely related, higher than a specified threshold. It creates an edge between variables
that are directly proportional to the point P of the connections. Please be aware that density vectors are connected via boundaries. The
issue of recognizing the Twitter support group as an exploding event with several bursting features for each grouping is the burst
features in bursting events. To find the grouping in the Twitter connection, it utilizes the Louvain technique.
It defines a hot occurrence as one that occurs over a short amount of time and generates a large amount of Twitter activity. It
estimates the event’s hotness degree using the list of bursting characteristics fi and their corresponding Twitter quantities for each
occurrence E. It identifies an occurrence with positive and negative quantifiers by integrating public responses to it. The hotness of the
function is determined using Eq. (5).
∑
m
vi
He (t) = snt (5)
t=1
t

The hotness level is denoted asHe (t), for a subject with m bursting occurrences, Twitter quantity vi and negativity profiles snt . The
time is denoted as t.
Fig. 2 shows the pictorial representation of He (t). The quality of the social media texts and the negativity profiles are used to
calculate He (t). The research study serves as the foundation for the suggested knowledge administration strategy for leveraging
massive social media information. Organizations may extract interesting information from social media by merging big data tech­
nologies, social media, and information sharing to uncover potential concerns, challenges, opportunities, and best practices. Previ­
ously, extracting information from massive social media information was confined to a small number of qualified professionals.
Thanks to integrated information management solutions, it is now possible to be used and benefited by both administrators and
staff. There are no actual recommendations or concepts for implementing an effective KM system and its working process in its

Fig. 2. Pictorial representation of He (t).

6
H. Zhang et al. Information Processing and Management 59 (2022) 102762

uncovered research.
Fig. 3 shows the proposed Big data-assisted social media analytics schematic view for business (BD-SMAB) model. It analyses social
media data and gives the business impact based on the analysis. It suggests an organizational learning architecture to assist interested
firms in leveraging big data analytics to meaningfully match social media information to opponents and uncover data analytics. There
is no structure in place in the KM space to assist businesses in leveraging massive social media information for business decision-
making.
In the old days, due to privacy and confidentiality issues and competitive market worries, it is generally difficult for organizations
to compare their actual information with competing companies. Even so, context-dependent methods trying to compare social media
information toward competing companies are now possible even though social media information is publicly ready for review.
Companies could benefit from understanding consumers’ thoughts in real-time by exploiting huge social media information and
uncovering opportunities, challenges, and concerns, such as recognizing whether consumer mood on social networks is getting
increasingly unfavorable. Combining huge social media information and knowledge management opens up a new route for firms to
gain significant business insight. It could significantly impact firms by providing critical customer service and enhancing their
competitive positioning analysis.
Big data technologies are employed as a solution in the suggested framework. To analyze large amounts of social media information
relating to organizations and their opponents and display and benchmark competition evaluations across activities, goods, problems,
and other domains that may affect operational efficiency.
Big data systems from IBM, Sybase, Mysql, and Windows can be combined to collect, organize, analyze, and information from
various Facebook pages. Massive data analysis and related information from big social media information facilitate such analyses.
Regularly, collected knowledge can be saved in a knowledge base and shared with top management via the organization’s current
management information system. This suggested technique is particularly appropriate to client service-oriented industries such as
finance, medical, retail, and e-commerce.
Customer service is critical for organizations in various service industries since it can help them stand out from their competition.

Fig. 3. Schematic view of the proposed Big data-assisted social media analytics for business (BD-SMAB) model.

7
H. Zhang et al. Information Processing and Management 59 (2022) 102762

Maintaining good client connections and establishing a profitable business requires excellent customer service and understanding. The
suggested architecture can improve a corporation’s client service, leading to greater business operations by collecting, preserving, and
applying derived consumer knowledge from massive social media information. It can explain a test case that we just did to illustrate the
utility of the proposed model.

3.1. Case study

Businesses want to know how to extract commercial importance from the millions of social media postings and communications
that customers send each day all over the world. It was discovered that social news media "earned" worth can have a long-term effect on
sales. While the "earned" worth of social media is dependent on the type of understanding it can recover from large amounts of social
information sources, the uniqueness and value of the suggested framework are also dependent on the derived social media information.
Over the last few decades, one of the writers has collaborated with VOZIQ ([Link] a social media analysis or­
ganization, to investigate extracted information from massive social media information. VOZIQ uses social media data analysis to
extract useful information from Twitter, Linkedin, blogging, newsgroups, and comment sections. To demonstrate the applicability of
the suggested methodology, it collaborated with VOZIQ on a case study involving analyzing 882,371 tweets related to five significant
retail businesses to generate new knowledge.
The VOZIQ social media monitoring technology was created with a concept similar to the suggested framework and has already
been used by several clients. Twitter, a major social media website, allows users to communicate and find subjects of interest in the real
moment with a community of "friends" by allowing them to post and receive brief messages called tweets. It may acquire comments
(texts) for selected organizations using the Twitter searching API.
It used various natural language processing such as textual data and sentiment classification to clean, organize, and enrich this data
set to assess sentiment generally and within subcategories. Around these businesses, it also spots new trends and topical themes. It
ables to detect deficiencies in certain products or services by analyzing the results obtained from these businesses. Leximancer was
chosen because it is a famous text analysis tool with a simple and intuitive user experience.
The map in the center indicates the relevance of the notions. It conducted the same study and created the same numbers for the
other four organizations to uncover hidden truths and make comparisons. It also analyzed the comments over the period, week by
week and monthly, to look for similarities and predict future events. It constructed a series of insight summaries to record the collected
information at the end of the research study. The following information has been derived:

• Voice Share: VOZIQ evaluates commentary from several firms within the same industry on social media. For reference, this
research study had contrasted comments from social media, which refer to Costco, eBay, Kohl, The Hardware Depot, and
Amazon, one or all of the aforementioned retail companies. It can observe that for the whole research time, Walmart was cited
most. An analysis of references to social media can assist in comprehending more clearly the customers’ attention for several
firms or brands in the same sector.
• Subjects and topics Sense: VOZIQ identifies and contrasts important themes of each company and their feelings. Substantial
phrases taken from references are the major way to identify the key concepts in the material.

VOZIQ may identify social media references according to the sorts of feeling into broad sections. Finally, computerized text analysis
is performed in each area to analyze the text (favorable neutral, unfavorable), find a certain community, and then do sentimental
analyses to determine feelings for each subcategory.
In this case, "cash card" gives Costco a pretty favorable feeling without any negative feeling, whereas "special offers," presumably
dependent on accessibility or tastes, produce a positive and negatory feeling. These topics give useful information for organizations to
enhance their offers, manage problems, and achieve success.

• 3C Emotion Benchmarking Report: VOZIQ provides social media information from its clients, rivals, and clients of the firm
named 3C monitoring to generate this summary for any enterprise. VOZIQ initially identifies these social media references into
major categories based on current target market intelligence. A computer algorithm is run to read and include relevant terms in
each division of each organization. The sentiment classification is carried out to gain the feeling rating for each area. This strategy
helps enterprises to find competitive information that works.

Costco (C) contrasts in sentiments across product, planning, and sales areas with four other firms, namely Kohls(K), Walmart(W),
Kmart(M), and The Habitat Depot(H). Peer rankings and categories are shown how one firm stacks other types of businesses. Drilldown
data at the segment level sends this study very workable because any firm can precisely know where it is dominating and where it must
be focused on becoming a champion in its particular market.

• The category structures: VOZIQ customizes and optimizes unmonitored and controlled categorization approaches to classes
based on a hybrid methodology. For instance, it has offered many Costco theme mapping from references to social media. A list of
groups is provided through the clustering analysis. A subject can be considered as a classification for each group. The highest

8
H. Zhang et al. Information Processing and Management 59 (2022) 102762

phrase recurrence decides the subject, and bigger rings reflect higher phrase frequency. It may be assumed that food products for
Costco were the most often mentioned issue.

The use of various colors in distinct groups is another significant element of this mapping. The same color is applied to groups of a
comparable category. The incidence of the references in groups of a similar kind may be readily shown using this function. The subjects
in green rings represent the domain Cost & Marketing, and the subject’ deals’ is mentioned most often concerning other subjects in the
domain Cost and Marketing.

• Correlation coefficients of classification: VOZIQ also detects connections between other subcategories to allow users to uncover
pairings of brand/product subjects and how consumers on social media are linked with each other. The correlations values were
calculated for each category pair in the test case. When the correlations were greater, it drew longer arc lines between groups. In
the favorable subject map, "Gift Certificates" and "Freebies" are strongly connected. The connections section on these maps can
assist firms in uncovering related areas and possible drivers of risks and solutions.

4. Software analysis and performance evaluation

The study framework for a collection of FMCG firms is company fan pages on Social media. These firms are available in all major
food chains, and their identities are strongly involved in social media. To connect and communicate with their customers, FMCG firms
have established their Facebook groups. Facebook is a company’s only social media tool for engaging customers. It gives an un­
equivocal and detailed framework for examining the influence of paid or controlled social media engagements on brand purchasing.
The corporation on Facebook has tracked each participant household panelist and the branded buy of the six brands. All six trademarks
are on the Social Media Page of the firm. The goods group in the marks is the same but with various brands (soft drinks or sweets).
Fig. 4(a) and 4(b) shows the social media message analysis and the social media mention analysis of the proposed Big data-assisted
social media analytics for business (BD-SMAB) model, respectively. Five different companies are considered for the simulation
analysis. The respective company mentions on Twitter are analyzed and plotted in the above figure. The same way the social media
mention of the Walmart company is plotted in the above figure. The results show the variations of the company mentions in social
media.
Table 1 shows the social media comments analysis of the proposed Big data-assisted social media analytics model for business (BD-
SMAB). Costco company is taken for the simulation analysis. The different products and services provided by the company, such as
clothing, accessories, customer service, parking, etc., are analyzed, and the positive comments and the mean of the comments are
analyzed from the social media are evaluated and tabulated in the above table. Over which jewelry selling has the highest positive
comments on the social media.
Fig. 5(a) and 5(b) shows the Twitter and Facebook stock index analysis of the proposed Big data-assisted social media analytics for
business (BD-SMAB) model, respectively. The social media index is analyzed daily in Twitter and Facebook, and the respective social
media mention of the company is monitored and plotted in the above figures. The social media index is higher for the Facebook
messages than the Twitter messages over the entire simulation. The performance of the proposed Big data-assisted social media an­
alytics for business (BD-SMAB) model has the highest accuracy.

Fig. 4. (a). Social media message analysis, Fig. 4(b). Social media mention analysis.

9
H. Zhang et al. Information Processing and Management 59 (2022) 102762

Table 1
Social media comment analysis of the proposed Big data-assisted social media analytics for business (BD-SMAB) model.
r Positive comment (%) Mean (%)

Products Clothing 57 63
Electronics 68 73
Jewelry 91 87
Watches 56 56
Services Parking 39 40
Returns and exchange 61 80
Customer service 51 37

Fig. 5. (a). Twitter stock index analysis, Fig. 5(b). Facebook stock index analysis.

Table 2 shows the social media mention analysis of the Big data-assisted social media analytics for business (BD-SMAB) model. Five
different companies are considered for the simulation analysis, and their social media mentions on Facebook and Twitter are analyzed
and tabulated in the above table. The Wal-Mart company has the highest mention, and the next highest is Costco. These companies
have the highest purchases over the month compare to the other companies. Where Kmart and Kohl’s company has the lowest social
media mentions in the entire month.
Fig. 6(a) and 6(b) shows the Facebook message frequency analysis and the brand purchase frequency analysis of the proposed Big
data-assisted social media analytics for business (BD-SMAB) model, respectively. Ten brands are considered for simulation, and the
respective brand mentions on Facebook and the number of purchases from the same brand are analyzed and plotted in the above
figures. It shows that the medium mentioned brands have a medium purchasing history than the highest and lowest social media
mentioned brands.
The proposed Big data-assisted social media analytics for business (BD-SMAB) model is designed and analyzed. This research
analyzes the different simulation parameters, such as social media mention, social media index, brand purchase, etc., are analyzed in
this research. The results show that the proposed Big data-assisted social media analytics for business (BD-SMAB) model has the
highest accuracy in analyzing social media and helps companies to improve their efficiencies.

Table 2
Social media mention analysis.
Company Social media mention (%)

Wal-Mart 35
Costco 24
Home depot 19
Kmart 11
Kohl’s 11

10
H. Zhang et al. Information Processing and Management 59 (2022) 102762

Fig. 6. (a). Facebook message frequency analysis, Fig. 6(b). Brand purchase frequency analysis.

5. Conclusion and future scope

In almost all domains of modern civilization, big data plays a crucial role. Big data analytics is currently employed in various
disciplines and industries, including medicine, economics, commerce, education, marketing, etc. New data is displayed by analyzing
the data produced daily. The latter can increase decision-making, improve the quality of decisions, and bring firms a competitive edge.
The research aims to assist decision-makers in using Big Data to enhance their organizational performance. It strives to strengthen the
organization’s decision-making processes by presenting the Big data-assisted social media analytics for business (BD-SMAB) model,
which combines Big Data and analytics into the decision-making stages. The initial study outcomes create a system that integrates the
decision-making mechanism with big data and analytics, and modeling. Models of choice are the most important contributions of the
work. The suggested paradigm begins with choice comprehension and modeling. The analytical approach in social media refers to
collecting and evaluating social media sites and blogs for decision-making. This technique goes beyond the normal monitoring or
fundamental analysis of retweets or "will" to establish a detailed concept of the social consumer.
This framework can guide decision-makers and the big data analysis group in the following steps. Thus, the modeling of decisions
using BD-SMAB enables coordination and teamwork between decision-makers and the predictive analytics group, clearly showing the
choice and needs. This approach continues to be a conceptual model based on past research findings and aims to assess and improve the
framework for future research. The proposed method examines social media analysis impacts on different areas such as real estate,
organizations, and beauty trade fairs. This diversity of these companies shows the effects of social media and how positive decisions
can be developed. Take better marketing decisions and develop a strategic approach. As a result, the BD-SMAB method enhance
customer satisfaction and experience and develop brand awareness.

CRediT authorship contribution statement

Honglei Zhang: Conceptualization, Visualization. Zhenbo Zang: Conceptualization, Visualization. Hongjun Zhu: . M. Irfan
Uddin: . M. Asim Amin: Formal analysis, Data curation.

Acknowledgements

The authors thank the support of Chongqing Social Science Planning PhD Program (Grant No:2016BS115); the support of Hu­
manities and Social Sciences Research Planning Project of Chongqing Education Commission in 2017 (Grant No: 17SKG141); the
support of the Science and Technology Research Program of Chongqing Municipal Education Commission (Grant No: KJ1709235).

References

Al Shahrani, H., & Ghandour, R. (2018). Impact of social media on purchase decision making of Saudi students in London. 12(1), 20-20.
Alhakimi, W., & Alwadhan, R. (2021). Social media and the decision-making process: Empirical evidence from Yemen. International Journal of Electronic Marketing and
Retailing, 12(2), 133–155.
Al-Turjman, F., & Salama, R. (2020). Security in social networks. Security in IoT Social Networks.
Balaanand, M., Karthikeyan, N., Karthik, S., Varatharajan, R., Manogaran, G., & Sivaparthipan, C. B. (2019). An enhanced graph-based semi-supervised learning
algorithm to detect fake users on Twitter. The Journal of Supercomputing, 75(9), 6085–6105.

11
H. Zhang et al. Information Processing and Management 59 (2022) 102762

Bhandari, R. S., & Bansal, S. (2019). An analysis between search engine optimization versus social media marketing affecting individual marketer’s decision–making
behavior. Jindal Journal of Business Research, 8(1), 78–91.
Chandrasekaran, G., Nguyen, T. N., & Hemanth, D. J. (2021). Multimodal sentimental analysis for social media applications: A comprehensive review (p. e1415). Wiley
Interdisciplinary Reviews: Data Mining and Knowledge Discovery.
Chang, Y. C., Ku, C. H., & Chen, C. H. (2019). Social media analytics: Extracting and visualizing Hilton hotel ratings and reviews from TripAdvisor. International
Journal of Information Management, 48, 263–279.
Gao, J., Wang, H., & Shen, H. (2020). Task failure prediction in cloud data centers using deep learning. IEEE Transactions on Services Computing, 32, 1111–1116.
Ghani, N. A., Hamid, S., Hashem, I. A. T., & Ahmed, E. (2019). Social media big data analytics: A survey. Computers in Human Behavior, 101, 417–428.
Gheisari, M., Najafabadi, H. E., Alzubi, J. A., Gao, J., Wang, G., Abbasi, A. A., et al. (2021a). OBPP: An ontology-based framework for privacy-preserving in IoT-based
smart city. Future Generation Computer Systems, 123, 1–13.
Gheisari, M., Najafabadi, H. E., Alzubi, J. A., Gao, J., Wang, G., Abbasi, A. A., et al. (2021b). OBPP: An ontology-based framework for privacy-preserving in IoT-based
smart city. Future Generation Computer Systems, 123, 1–13.
Golpayegani, S. A. H., Esmaeili, L., Mardani, S., Mutallebi, S. M., & A survey of trust in social commerce. (2019). E-systems for the 21st century: Concept, developments,
and applications-two volume set (p. 1). ACM Computing Surveys.
Hansen, J. M., Saridakis, G., & Benson, V. (2018). Risk, trust, and the interaction of perceived ease of use and behavioral control in predicting consumers’ use of social
media for transactions. Computers in Human Behavior, 80, 197–206.
Hossain, M. A., Dwivedi, Y. K., Chan, C., Standing, C., & Olanrewaju, A. S. (2018). Sharing political content in online social media: A planned and unplanned behavior
approach. Information Systems Frontiers, 20(3), 485–501.
Kovacova, M., Kliestik, T., Pera, A., Grecu, I., & Grecu, G. (2019). Big data governance of automated algorithmic decision-making processes. Review of Contemporary
Philosophy, 18, 126–132.
Kumar, G., Saha, R., Buchanan, W. J., Geetha, G., Thomas, R., Rai, M. K., et al. (2020). Decentralized accessibility of e-commerce products through blockchain
technology. Sustainable Cities and Society, 62, Article 102361.
Liang, T. P., & Liu, Y. H. (2018). Research landscape of business intelligence and big data analytics: A bibliometrics study. Expert Systems with Applications, 111, 2–10.
Ma, C., Au, N., & Ren, L. (2020). Biased minds experience improved decision-making speed and confidence on social media: A heuristic approach. Information
Technology & Tourism, 22(4), 593–624.
Ngan, R. T., Ali, M., Fujita, H., Abdel-Basset, M., Giang, N. L., Manogaran, G., et al. (2019). A new representation of intuitionistic fuzzy systems and their applications
in critical decision making. IEEE Intelligent Systems, 35(1), 6–17.
Nie, X., Fan, T., Wang, B., Li, Z., Shankar, A., & Manickam, A. (2020). Big Data analytics and IoT in operation safety management in under water management.
Computer Communications, 154, 188–196.
Noureddine, O. H., & ZeinEddine, R. B. (2018). Social media and its impression on consumers’ behavior during their decision-making process. American Scientific
Research Journal for Engineering, Technology, and Sciences (ASRJETS), 41(1), 76–84.
Oksana, B. A. D. Z. I. U. N. (2020). The role of social media in various travel decision-making stages. Journal of Social Sciences and Humanities Research, 8(1).
Pham, D. V., Nguyen, G. L., Nguyen, T. N., Pham, C. V., & Nguyen, A. V. (2020). Multi-topic misinformation blocking with budget constraint on online social
networks. IEEE Access: Practical Innovations, Open Solutions, 8, 78879–78889.
Song, Z., Sun, Y., Wan, J., Huang, L., Xu, Y., & Hsu, C. H. (2019). Exploring robustness management of social internet of things for customization manufacturing.
Future Generation Computer Systems, 92, 846–856.
Tafesse, W., & Wien, A. (2018). Implementing social media marketing strategically: An empirical assessment. Journal of Marketing Management, 34(9–10), 732–749.
Voramontri, D., & Klieb, L. (2019). Impact of social media on consumer behavior. International Journal of Information and Decision Sciences, 11(3), 209–233.
Yang, P., Yang, Y., Wang, Y., Gao, J., Sui, N., Chi, X., et al. (2016). Spontaneous emission of semiconductor quantum dots in inverse opal SiO2 photonic crystals at
different temperatures. Luminescence: The Journal of Biological and Chemical Luminescence, 31(1), 4–7.
Zhao, Y., Yu, Y., Shakeel, P. M., & Montenegro-Marin, C. E. (2021). Research on operational research-based financial model based on e-commerce platform.
Information Systems and e-Business Management, 1–17.
Zong, K., Yuan, Y., Montenegro-Marin, C. E., & Kadry, S. N. (2021). Or-based intelligent decision support system for e-commerce. Journal of Theoretical and Applied
Electronic Commerce Research, 16(4), 1150–1164.

12