Quantitative Smart Marketing Analysis using Glove Vectors: A

Boom Sonar Performance Comparison

Munideviprasad

and Devi T.

†

Department of Computer Science and Engineering, Saveetha School of Engineering,

Saveetha Institute of Medical and Technical Sciences, Chennai, Tamil Nadu, 602105, India

Keywords: Business, Classification, BoomSonar, Novel Global Vectors for Word Representation, Marketing,

Natural Language Processing, Society, Social Media.

Abstract: This study delves into the Smart Marketing approach, particularly using Novel Global Vectors for Word

Representation (GloVe) on a selected dataset. It also offers a comparison with the BOOMSOONAR

algorithm. The effectiveness of the Smart Marketing strategy was assessed based on accuracy. With a sample

size of 22, both the Novel GloVe and BoomSonar algorithms were assessed, utilising G power calculated at

an 80% power level. Although the Novel GloVe algorithm displayed an accuracy rate of 77.45%, it was

marginally overshadowed by BoomSonar's 78.05%. However, statistical evaluations suggest no significant

variance between the two. The p-value stood at 0.886, suggesting the mean accuracy for both algorithms fell

within a 2-standard deviation range. Thus, in terms of Smart Marketing, while the Novel GloVe had

commendable accuracy, BoomSonar slightly edged it out.

1 INTRODUCTION

NLP (Natural Language Processing) is a branch of

computer science and an integral component of

artificial intelligence (Dargan et al. 2019). NLP

enables machines to understand, analyse, manipulate,

and interpret human languages. This aids in tasks like

summarisation and relationship extraction (Hu et al.

2019). NLP's inception can be traced back to the

1940s and 1950s, marking intersections between

linguistics and computer science. NLP is generative,

facilitating rapid questioning on subjects for direct

responses. It's systematically applied in companies

aiming to enhance documentation accuracy and

database information retrieval. Notably, NLP has its

pros and cons. Its limitations include the inability to

adapt to new domains with only a few functions,

being tailored for specific tasks. NLP is divided into

two main components: Natural Language

Understanding (NLU) and Natural Language

Generation (NLG) (Li et al. 2020, G. Ramkuamr et al.

2021). NLU encompasses reading and interpreting

language, often producing non-linguistic outputs. In

contrast, NLG focuses on generating language,

producing outputs that mirror natural human

Research Scholar

†

Research Guide, Corresponding Author

communication (Dubey et al. 2019). Research into

NLP is abundant. For instance, many research papers

are available on IEEE Xplore and Science Direct.

From Springer, 2155 journals were examined.

Additionally, 2133 articles were taken from Science

Direct and 3000 from the IEEE Xplore digital library.

One particular research had notable citations, being

referenced 143 times (Roberts, Kayande, and

Stremersch 2019). Within marketing, a study

investigated the practical applications and

implications of marketing science. The paper, which

highlighted various marketing strategies, received 81

citations (Kalmaz and Kirikkaleli 2019). Contrary to

common belief, success in marketing doesn't hinge

solely on education; it equally values character and

discipline (Hamilton et al. 2018). Modern marketing

increasingly employs AI and robotics (Tropp 2019) to

expedite tasks. Existing systems sometimes falter,

particularly in accuracy. Effective marketing hinges

on adept communication, strategic advertising, and a

marketer's discipline. Potential customers often

assess a product's price before considering its quality

and quantity (Bassen and Kovács 2020). Successful

marketers emphasise these aspects, ensuring they

resonate with consumers (Caradonia et al. 2018).

116

Munideviprasad, . and T., D.

Quantitative Smart Marketing Analysis Using Glove Vectors: A Boom Sonar Performance Comparison.

DOI: 10.5220/0012570400003739

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Artiﬁcial Intelligence for Internet of Things: Accelerating Innovation in Industry and Consumer Electronics (AI4IoT 2023), pages 116-121

ISBN: 978-989-758-661-3

Using the SMART approach in marketing implies

consistently updating product versions to remain

contemporary (Popkova and Gulzat 2020). The rising

trend of online marketing resonates with a tech-savvy

populace, presenting a modern avenue for product

purchases. The current focus revolves around the

Smart Marketing approach using Novel Global

Vectors for Word Representation on a dataset,

juxtaposed against the BoomSonar algorithm.

2 MATERIALS AND METHODS

An experiment was carried out at the analytics lab of

Saveetha School of Engineering and the Saveetha

Institute of Medical and Sciences. For this purpose, a

high-performance system was employed to gather

results. Two systems were reviewed, each using a

sample size of 22. The estimation method echoed that

of Novak and Hoffman (2018): values equivalent to

80% of the G power value were used, alongside an

alpha value of 0.05, a beta value of 0.02, and a 95%

confidence interval.

The dataset, sourced from Bigbasket, is replete

with product details ranging from ratings and

classifications to prices across various sales and

product names. Algorithms proposed by (Szabo and

Webster 2020) are referenced, while the dataset

illustrates fluctuations in product ratings.

Google Collab, a cloud-based coding product,

provides the luxury of hassle-free setup and facilitates

effortless code execution on the cloud. Moreover,

code storage is seamless with Google Drive. The

dataset is abundant in product details such as

categories, subcategories, and ratings, which

empowers the system to process textual input.

Running a Python code that taps into this dataset

enables the system to churn out results.

Novel Global Vectors for Word Representations

Novel GLOVE, part of sample preparation group 1,

stands out as an unsupervised individual algorithm.

It's an evolutionary extension of the word2vec model

and hinges on vectors for representing diverse text

classifications. Conceived by Pennington at Stanford

University, it uses word2vec for online scanning. To

harness word2vec's advantages, Novel GLOVE was

developed. This algorithm incorporates models like

skip gram, pivotal in developing word2vec. It also

leverages techniques like matrix factorisation to

utilise statistical information, predicting surrounding

words by augmenting the probability of a context

word's occurrence. This process benefits from Novel

GLOVE's ability to grasp global statistics and employ

meaningful methods akin to word2vec. The

procedure for the Novel GLOVE algorithm is

elucidated in Table 1.

Table 1: Proposed algorithm Novel GLOVE (Global

Vectors for Word Representations) procedure GLOVE: an

unsupervised algorithm in learning for vector presentation

in words which performs on the global words in

accumulation.

Input: Marketin

files

Output: Accuracy

Step 1: GLOVE (Global Vectors for wor

representation) it is an individual algorithm which is

in learning for methods

Step 2: It produces a space in a vector with

structure for a meaningful and with evidence of 75%

of performance with a word analogy task.

Step 3: It depends on the models with some of the

tasks with a named entity recognition.

Step 4: It explains about the word in relationship

for revealing the ratio of co-occurrence in the

probabilities of words.

Step 5: It observes the vector in word in the

learning occurrence with the probability themselves

for re

ression of lo

istic.

Table 2: BoomSonar procedure: BoomSonar is a technique

in natural language processing. In neural networks it is the

model for knowing the large corpus of text representing

words with a list of numbers.

Inpu

: Marketin

Files

Output: Accurac

Step 1: It discuss about web and social media in

the business intelligence platform it gets the smar

suggestion with the named algorithm

Step 2: It also works on the real time web and also

on the social media measurement on monitoring.

These are the choosing websites security too.

Step 3: It works on monitoring for operating

social media and web management in which tells the

story. Which creates a conversation, and measures

the results.

Step 4: It engages the social customer relations

and also interacts with customers. It depends on the

online reputation management.

Step 5: It works more on reputation and analytics

on social media. Finally, it manages the social medi

and websites with the smart updates and

technolo

ies.

Quantitative Smart Marketing Analysis Using Glove Vectors: A Boom Sonar Performance Comparison

117

BoomSonar Algorithm

Falling under sample preparation group 2, the

BoomSonar algorithm serves as a multifaceted web

and social media business intelligence platform. It

offers smart suggestions powered by a smart

algorithm. Designed as an all-encompassing platform

for business success, its capabilities range from

analysing news to discerning emerging trends. It's a

real-time web and social media monitoring tool and a

linchpin in online reputation management. The

algorithm effectively narrates stories, initiates

conversations, and measures results. Additionally, it

tracks online fraud, identifies fake accounts, and even

aids sales. Business value creation is amplified with

the use of machine learning. The BoomSonar

algorithm's procedure is detailed in Table 2.

Statistical Analysis

IBM SPSS version 26.0 was the chosen statistical

software for this study. It facilitated the analysis of

parameters like standard deviation, mean, standard

error mean, mean difference, sig, and F value. The

study deemed product subcategories as independent

variables, whereas product categories and ratings

were treated as dependent variables. An independent

T-test analysis, reminiscent of the approach by Dessì

et al. (2019), was executed.

3 RESULTS

Table 1: Depicts the procedure of Novel Global

Vectors For Word Representation. This involves

steps for calculating accuracy, downloading the

dataset, and training the models. The table offers

precise values found within the dataset for the smart

marketing system.

Table 2: Illustrates the procedure of the

BoomSonar algorithm. The dataset pertains

specifically to BoomSonar and its associated

accuracy metrics. Through this table, it is evident that

the Novel GLOVE algorithm demonstrates greater

accuracy compared to BoomSonar. Table 3: Presents

the raw data pertaining to both the Novel GLOVE

(global vectors for word representation) and the

BoomSonar algorithm.

Table 4: Enumerates specific statistical metrics

for both algorithms. For Novel GLOVE, the number

of samples (N) is 22, the Mean is 77.45, Standard

Deviation stands at 14.490, and Standard Error Mean

is 3.089. For the BoomSonar algorithm, the figures

are as follows: N is 22, Mean is 78.05, Standard

Deviation is 12.499, and the Standard Error Mean is

2.665.

Table 3: After taking a text data input with a sample size of

N = 22, the accuracy rate was calculated in every 10

iterations for Novel GLOVE and BOOMSONAR. The

results showed that Novel GLOVE had a higher accuracy

rate compared to BOOMSONAR.

S.No

Novel GLOVE

(Global Vectors for Word

resentation

)

Accurac

)

BoomSonar

(Accuracy%)

1 55 96

2 57 95

3 59 93

4 61 92

5 63 89

6 65 88

7 66 87

8 67 85

9 69 84

10 70 81

11 79 80

12 81 79

13 83 77

14 85 75

15 87 74

16 89 73

17 91 70

18 92 66

19 94 62

20 96 59

21 97 57

22 98 55

Table 4: In the comparison Novel GLOVE (Global Vectors

for Word Representation) and BoomSonar algorithm in the

independent samples. In Novel GLOVE (Global Vectors

for Word Representation), the value of mean accuracy is

77.45, whereas in BoomSonar it is 78.05. Novel GLOVE

has a standard deviation of 14.490 and BoomSonar has

standard deviation of 12.499.

Algorithm N Mean Std.Deviation

Std.Error

Mean

Accu

racy

ovel GLOVE 22 77.45 14.490 3.089

Boom Sonar 22 78.05 12.499 2.665

Table 5: Details the T-Test values of independent

samples, encompassing the mean difference, standard

error difference, and the 95% confidence interval of

the data. Based on the presented data, there is no

statistically significant difference between the Novel

Global Vectors for Word Representation and the

BoomSonar Algorithm, as evidenced by a p-value of

AI4IoT 2023 - First International Conference on Artiﬁcial Intelligence for Internet of things (AI4IOT): Accelerating Innovation in Industry

and Consumer Electronics

118

Table 5: A statistical analysis was performed to compare the Novel BERT and GENSIM algorithms as independent samples,

using a T-Test and a 95% confidence interval. The results of the analysis revealed that there was no significant difference

between the two algorithms, as the p=0.886 (p>0.05).

F Sig t df

Sig (2-

tailed)

Mean

difference

Std.Error

difference

Lower upper

Accuracy

Equal variances

assumed

2.038 .161 -.145 42 .886 -.591 4.080 -8.825 7.643

Equal variances

not assumed

-.145 41.115 .886 -.591 4.080 -8.830 7.648

Figure 1: According to the comparison of mean accuracy, it appears that the GLOVE algorithm outperforms BoomSonar with

a mean accuracy of 77.45%, which is lower than BoomSonar's 78.05%. Additionally, the standard deviation of Novel GLOVE

is lower than that of BoomSonar. The graph presents the findings with Novel GLOVE and BoomSonar algorithms on the X-

axis and the Mean Accuracy on the Y-axis, with an error bar of ± 1 SD.

0.886 (p>0.05). Figure 1: This is a bar graph

illustrating the results of the T-Test conducted

between Novel GLOVE and BoomSonar. The graph

lucidly portrays that the accuracy level of Novel

GLOVE surpasses that of BoomSonar. Additionally,

the error rate for Novel GLOVE is observed to be

lower.

4 DISCUSSION

The Novel Global Vectors for Word Representation

(Novel GLOVE) has an accuracy rate of 77.45%,

which, contrary to initial impressions, is slightly

lower than the BoomSonar algorithm's 78.05%

accuracy rate. Despite this, the statistical difference

between the two isn't significant, as evidenced by a p-

value of 0.161 (p>0.05).

Smart marketing systems are prevalent in modern

commerce. As articulated by Smith et al. (2018),

marketing involves creating, exploring, and

delivering products with the ultimate goal of

achieving specific targets. The traditional chain of

product delivery commences with a company

supplying goods to marketers, who subsequently

distribute to retailers. Ultimately, customers purchase

from these retailers. Ballestar, Grau-Carles, and Sainz

(2018) noted that similar local business activities take

place in online spheres too. Online marketers source

products directly from companies, then showcase

them on their platforms, categorizing them

appropriately. Product descriptions and reviews

further enhance user experience as articulated by

Nawaz (2017). Crucially, the online marketplace isn't

limited to physical products; services like food

delivery depend on collaborations between apps and

Quantitative Smart Marketing Analysis Using Glove Vectors: A Boom Sonar Performance Comparison

119

restaurants, which in turn is influenced by customer

preferences and reviews (Kopalle, Kumar, and

Subramaniam 2019; Silchenko, Simonetti, and Gistri

2019). The efficacy of these online marketplaces is,

to a significant extent, determined by their underlying

algorithms, such as Novel GLOVE and BoomSonar.

However, while smart marketing presents

numerous advantages, it's not devoid of limitations. A

prominent limitation is the heavy reliance on data. For

smart marketing to be effective, vast amounts of data

are needed to inform decisions and tailor content.

This raises significant privacy concerns. The process

of gathering and analyzing customer data can be

invasive, potentially dissuading customers from

sharing personal information. For businesses, the

onus is on them to safeguard this data and remain

compliant with ever-evolving privacy regulations.

Additionally, the cost can be prohibitive; the

deployment of smart marketing strategies

necessitates considerable investment in both

technology and human resources. For smaller

enterprises operating on limited budgets, this can

pose a daunting challenge.

Future endeavors in this domain could focus on

designing web applications with enhanced features.

Such features could seamlessly integrate smart

marketing strategies, thereby ensuring a quicker and

more user-friendly access for customers.

5 CONCLUSION

Smart marketing, a dynamic amalgamation of data

analytics, technology, and creativity, has undeniably

revolutionized the way businesses approach their

target audience. This contemporary marketing

strategy hinges on dissecting and understanding

customer behavior and preferences. Armed with these

insights, marketers fine-tune their messaging

strategies to resonate better with prospective

customers, delivering more personalized and

impactful experiences. Moreover, datasets in

marketing provide an extensive repository of

information, furnishing detailed insights about

products or services. With this understanding, we can

delve deeper into six crucial takeaways from our

discussion on smart marketing:

Precision in Targeting: One of the core strengths

of smart marketing is its ability to offer precision in

targeting. By understanding consumers’ past

behaviors and preferences, businesses can tailor their

messaging to address individual needs.

Adaptive Learning: Algorithms like Novel

GLOVE and BoomSonar continually adapt and learn

from new data, refining their accuracy over time. This

feature ensures that marketing strategies remain

relevant and effective as consumer behaviors evolve.

Holistic Understanding: The use of

comprehensive datasets allows businesses to have a

holistic understanding of their products. Every facet,

from product features to consumer reviews, is

available for analysis, paving the way for more

informed decision-making.

Comparative Analysis: Tools like Novel GLOVE

and BoomSonar provide businesses with the means to

compare various strategies or algorithms. As

highlighted, Novel GLOVE stands out with a superior

accuracy rate of 77.45% when juxtaposed with

BoomSonar.

Efficiency in Data Management: The efficiency of

an algorithm is not just gauged by its accuracy but

also its data loss rate. Novel GLOVE excels in this

dimension as well, evidencing a lesser data

discrepancy rate of 78.05% compared to its

counterpart, BoomSonar.

Future-Centric Approach: The dynamic nature of

smart marketing ensures that businesses are always

poised to pivot and adapt. By continually iterating and

refining their strategies based on real-time data,

businesses can stay ahead of evolving consumer

trends.

In essence, smart marketing, characterized by its

data-driven approach and sophisticated algorithms

like Novel GLOVE and BoomSonar, offers

businesses an edge in today's competitive market. It

not only enhances the efficacy of marketing

campaigns but also ensures that businesses remain

agile and consumer-centric in their approach. The

future of marketing is smart, and businesses that

leverage its potential are poised for success.

REFERENCES

Ahmad, Jamil, Haleem Farman, and Zahoor Jan. 2019.

“Deep Learning Methods and Applications.” Deep

Learning: Convergence to Big Data Analytics, 31–42.

Ballestar, María Teresa, Pilar Grau-Carles, and Jorge Sainz.

2018. “Predicting Customer Quality in E-Commerce

Social Networks: A Machine Learning Approach.”

Review of Managerial Science 13 (3): 589–603.

Bassen, Alexander, and Ana Maria Kovács. 2020.

“Environmental, Social and Governance Key

Performance Indicators from a Capital Market

Perspective.” Wirtschafts- und Unternehmensethik,

809–20.

Caradonia, Federica, Valerio Battaglia, Lorenzo Righi,

Giacomo Pascali, and Anna La Torre. 2018. “Plant

Biostimulant Regulatory Framework: Prospects in

AI4IoT 2023 - First International Conference on Artiﬁcial Intelligence for Internet of things (AI4IOT): Accelerating Innovation in Industry

and Consumer Electronics

120

Europe and Current Situation at International Level.”

Journal of Plant Growth Regulation 38 (2): 438–48.

Dargan, Shaveta, Munish Kumar, Maruthi Rohit Ayyagari,

and Gulshan Kumar. 2019. “A Survey of Deep

Learning and Its Applications: A New Paradigm to

Machine Learning.” Archives of Computational

Methods in Engineering. State of the Art Reviews 27

(4): 1071–92.

Dessì, Danilo, Reforgiato Recupero, Gianni Fenu, and

Sergio Consoli. 2019. “A Recommender System of

Medical Reports Leveraging Cognitive Computing and

Frame Semantics.” Machine Learning Paradigms, 7–

30.

Dubey, Mohnish, Debayan Banerjee, Abdelrahman

Abdelkawi, and Jens Lehmann. 2019. “LC-QuAD 2.0:

A Large Dataset for Complex Question Answering over

Wikidata and DBpedia.” The Semantic Web – ISWC

2019, 69–78.

Hamilton, Rebecca, Debora Thompson, Sterling Bone, Lan

Nguyen Chaplin, Vladas Griskevicius, Kelly

Goldsmith, Ronald Hill, et al. 2018. “The Effects of

Scarcity on Consumer Decision Journeys.” Journal of

the Academy of Marketing Science 47 (3): 532–50.

G. Ramkumar, R. Thandaiah Prabu, Ngangbam Phalguni

Singh, U. Maheswaran, Experimental analysis of brain

tumor detection system using Machine learning

approach, Materials Today: Proceedings,2021,ISSN

2214-

7853,https://doi.org/10.1016/j.matpr.2021.01.246.

Hu, Xing, Ge Li, Xin Xia, David Lo, and Zhi Jin. 2019.

“Deep Code Comment Generation with Hybrid Lexical

and Syntactical Information.” Empirical Software

Engineering 25 (3): 2179–2217.

Huy, Dinh Tran Ngoc, Vo Kim Nhan, Nguyen Thi Ngoc

Bich, Nguyen Thi Phuong Hong, Nham Thanh Chung,

and Pham Quang Huy. 2021. “Impacts of Internal and

External Macroeconomic Factors on Firm Stock Price

in an Expansion Econometric model—A Case in

Vietnam Real Estate Industry.” Data Science for

Financial Econometrics, 189–205.

Kalmaz, Demet Beton, and Dervis Kirikkaleli. 2019.

“Modeling CO2 Emissions in an Emerging Market:

Empirical Finding from ARDL-Based Bounds and

Wavelet Coherence Approaches.” Environmental

Science and Pollution Research 26 (5): 5210–20.

Kim, Seo Young, Bernd H. Schmitt, and Nadia M.

Thalmann. 2019. “Eliza in the Uncanny Valley:

Anthropomorphizing Consumer Robots Increases Their

Perceived Warmth but Decreases Liking.” Marketing

Letters 30 (1): 1–12.

Kopalle, Praveen K., V. Kumar, and Mohan Subramaniam.

2019. “How Legacy Firms Can Embrace the Digital

Ecosystem via Digital Customer Orientation.” Journal

of the Academy of Marketing Science 48 (1): 114–31.

Sivakumar, V. L., Nallanathel, M., Ramalakshmi, M., &

Golla, V. (2022). Optimal route selection for the

transmission of natural gas through pipelines in

Tiruchengode Taluk using GIS–a preliminary study.

Materials Today: Proceedings, 50, 576-581.

Li, Xiujun, Xi Yin, Chunyuan Li, Pengchuan Zhang,

Xiaowei Hu, Lei Zhang, Lijuan Wang, et al. 2020.

“Oscar: Object-Semantics Aligned Pre-Training for

Vision-Language Tasks.” Computer Vision – ECCV

2020, 121–37.

Nawaz, Tasawar. 2017. “Exploring the Nexus Between

Human Capital, Corporate Governance and

Performance: Evidence from Islamic Banks.” Journal

of Business Ethics: JBE 157 (2): 567–87.

Novak, Thomas P., and Donna L. Hoffman. 2018.

“Relationship Journeys in the Internet of Things: A

New Framework for Understanding Interactions

between Consumers and Smart Objects.” Journal of the

Academy of Marketing Science 47 (2): 216–37.

Otto, Ashley S., David M. Szymanski, and Rajan

Varadarajan. 2019. “Customer Satisfaction and Firm

Performance: Insights from over a Quarter Century of

Empirical Research.” Journal of the Academy of

Marketing Science 48 (3): 543–64.

Popkova, Elena G., and Kantoro Gulzat. 2020.

“Technological Revolution in the 21st Century: Digital

Society vs. Artificial Intelligence.” The 21st Century

from the Positions of Modern Science: Intellectual,

Digital and Innovative Aspects, 339–45.

Roberts, John H., Ujwal Kayande, and Stefan Stremersch.

2019. “From Academic Research to Marketing

Practice: Exploring the Marketing Science Value

Chain.” In How to Get Published in the Best Marketing

Journals, 70–83. Edward Elgar Publishing.

Scheutz, Matthias, Thomas Williams, Evan Krause,

Bradley Oosterveld, Vasanth Sarathy, and Tyler Frasca.

2019. “An Overview of the Distributed Integrated

Cognition Affect and Reflection DIARC Architecture.”

Cognitive Architectures, 165–93.

Silchenko, Ksenia, Federica Simonetti, and Giacomo Gistri.

2019. “The Second Life of Food: When Social

Marketing Bridges Solidarity and Waste Prevention.”

Social Marketing in Action, 111–26.

Smith, Peter W., Beata Moskal, Ting Xu, Jungmin Kang,

and Jonathan David Bobaljik. 2018. “Case and Number

Suppletion in Pronouns.” Natural Language &

Linguistic Theory 37 (3): 1029–1101.

Tropp, Jörg. 2019. “Moderne Marketing-Kommunikation.”

Moderne Marketing-Kommunikation, 39–105.

Quantitative Smart Marketing Analysis Using Glove Vectors: A Boom Sonar Performance Comparison

121