MACHINE LEARNING AND AI PLATFORMS FOR A

BUSINESS PROBLEM
 Table of Contents
Introduction......................................................................................................................................3

Background Study........................................................................................................................3

Problem Statement.......................................................................................................................3

Discussion........................................................................................................................................4

Applications of AI/Machine Learning in current operations of Automotive Fuel Organizations

in Australia...................................................................................................................................4

AI/Machine Learning Technique recommended for Automotive Fuel Organizations in

Australia.......................................................................................................................................5

Methodology....................................................................................................................................9

Predictive Analysis for modeling sample data- Automotive Fuel Organizations in Australia....9

Organizational Benefits..............................................................................................................12

Conclusion.....................................................................................................................................13

Reference List................................................................................................................................15

2
Introduction

Background Study
The advent of machine learning in recent times can be conceived as a novel development. This
can be accredited to the capability of providing machines with access to large amounts of data
and allowing them to perform complex mathematical calculations repeatedly while also learning
for themselves. The prospect of any business, ranging from banking to e-commerce from real
estate property to homeland security to universal health care to marketing in the stock market,
from manufacturing to education to retail to government organizations, is dependent on the
predictive analytics expertise that is developed and expanded in the near future. The rapidity
with which technology has changed in recent times has posed a significant challenging problem
encompassing all businesses. Technologies such as the Internet of Things, and the associated
features and capabilities of artificial intelligence (AI), machine learning (ML), and data analytics
are now available.  They are being steadily integrated into practical applications in order to
improve business intelligence (BI), and improve decision support with consequent faster and
augmented processing efficiency and the discovery of new technologies that increase the speed
with which vulnerabilities are identified, and generate significant cost savings and profitability
margins. 

Problem Statement
Automobile manufacturers can improve quality and output while also reducing waste when they
have access to the insights provided by big data. These are processes that are critical in today's
highly competitive market. In an increasingly data-driven environment, manufacturers are able to
solve problems early and deliver more agile business strategies, resulting in increased
productivity and profitability. The present report has been prepared to address the concerns
underlying applications of AI/Machine learning algorithms within automotive fuel industries
operative within Australia. It also underpins the benefits associated with the successful
implementation of AI/Machine learning techniques within the context of automotive industry
operations. Furthermore, the predictive analysis of the automotive industry sample data enables
the determination of the estimated profits or revenue and calculations pertaining to “Return on
Investment (ROIs).”

3
Discussion

Applications of AI/Machine Learning in current operations of Automotive Fuel

Organizations in Australia
Manufacturing operations underwent significant transitions as a result of the digitalization of
manufacturing and production processes. In addition to equipment conditions including
temperature and pressure, the AI system also collects data on human performance (Abduljabbar
et al., 2019). Aside from that, it keeps track of scrap customer delivery schedules, metal
conditions, and electric costs. The technology enables the mill to generate multiple grades of
steel by lowering operating costs, thereby resulting in competitive advantages in the marketplace.
In operations management, accuracy and efficiency are critical performance metrics. Artificial
intelligence presents the opportunity to improve productivity through increased efficiency while
also improving the precision associated with operations through increased accuracy. The use of
artificial intelligence (AI) in businesses can benefit them by addressing three primary needs:
"automating business processes, gaining insight through data analysis, and engaging with
customers and employees." The use of AI can also be applied to the development of algorithms
that can sort datasets and identify patterns in data, which is essential for business analytics. In the
case of the manufacturing industry, analytics practices are utilized to automate the
entire production process, optimize products, reduce costs by reducing energy waste, and
derive useful information by analyzing data collected from various aspects from across
the enterprises (Agrawal et al., 2019). Businesses benefit from the incorporation of new data
analysis and collection methods, including the existing product expansion or the development of
new digitalized products, because it allows them to generate product usage information and
thereby improve their products to better meet the demands of the customers. In the second place,
there is the integration and digitalization of the corresponding vertical and lateral value chains.
Examples of these aspects include manufacturing equipment, labour activity, manufacturing
process, and assessment of environmental conditions. Manufacturing is highly dependent on
technology and machinery in order to maintain high levels of efficiency and safety while
producing. Artificial intelligence (AI) is increasingly likely to be found in supply chain and
administration areas in the manufacturing industry so then it is to be utilized particularly for
production objectives within the industry. In this regard, it becomes imperative to mention that

4
advanced analytics practices make use of algorithms to map out the optimal operating condition.
Algorithms acquire the data and analyze it (Attaran and Deb, 2019). Workers located on the job
site use portable devices to monitor product quality, model solutions to issues, and share data
with colleagues across other locations, among other things. The “Blockchain” technology
implemented, on the contrary, has the potential to be applied across the entire manufacturing
industry in order to address a wide variety of projects as well as stakeholders.

Engineering and design professionals in the manufacturing industry can benefit from artificial
intelligence (AI). For example, "generative design" is a technique in which a user enters a
proposed design into a software program, and the program then generates possible solutions as
well as potential finished product concepts. AI can be further incorporated for a variety of other
tasks, such as developing predictions about interactions across production units, studying and
comparing relationships existing between different components of the manufacturing process
operating in tandem together, and enabling automated requests to provide updated information
on necessary components, tools, maintenance, and labour for manufacturing equipment
(Balashanmugam, 2021). The improved efficiency that AI offers can provide for the overall
improvement of the employees throughout the entire manufacturing process leading to
dramatically increased productivity that can be provided by AI for its customers.

AI/Machine Learning Technique recommended for Automotive Fuel Organizations

in Australia
A subset of artificial intelligence, machine learning (ML) includes the techniques that allow
computers to discern underlying patterns out from data and effectively deliver AI applications. A
key aspect of the machine learning algorithm is its iterative nature. This is essential since models
are increasingly exposed to new data; they are capable of adapting independently to the new
environment (Dagnaw, 2020). They learn from their previous computations and use this
knowledge to make decisions and deliver results that are repeatable and reliable. Artificial
Intelligence (AI) or Machine Learning (ML) is an amalgamation of several types that can be
classified into four principal categories namely: 1) Supervised, 2) Semi-Supervised, 3)
Unsupervised and eventually 4) Reinforced types. There are several machine learning algorithms
that are utilized in the industrial sector. They could be classified as either unsupervised or
supervised, which is largely dependent on the fact as to how they learn and derive essential

5
information from within the data. The following research report has been prepared by keeping
hindsight of the implementation of the surveyed ML algorithms within the automobile industrial
sector of Australia.

Starting with the basics, Statistical Learning Theory (SLT) is an industry-standard supervised
machine learning framework. The greatest advantage of SLT is the wide range of strategies and
scenarios in which it can be used. It is also entirely feasible to get around the problem of
observer variability by utilizing SLT methods (Hyder et al., 2018). The main goal is to provide
the best estimate of the output for inputs that have not yet been seen. Although a large number of
samples are required for this task, the application of SLT in some cases may result in over-
fitting. BNs, also known as the Bayesian Network, are an example of an SLT implementation.
BNs are used to describe the probability relationships existing between a number of different
variables. NBNs, which are amongst the simplest type of Bayesian network models, and are
similar to BNs. Based on theoretical considerations, NBNs could be summarized in the following
way: The naive Bayesian classifier realizes the conditional probability Ai corresponding to each
attribute of the data (training data) when allotted a class label C. It is then necessary to apply
classification, which is accomplished by employing the Bayes rule to calculate the probability of
C for each instance of A1, An, and then making a prediction regarding the highest class based on
the posterior probability (Johnston, 2021). This evaluation is generally predicated on the
assumption that all of the Ai attributes are conditionally independent when the value of class C is
taken into consideration. The corresponding mathematical formulation is depicted as follows

(1.)

Whenever Pr(C) > 0 for any possible combination of A, B, and C, there is a problem.

 The schematic diagram outlining the proposed structure of AI Architecture implementation

within the automotive industry is depicted as follows

6
 Figure 1: The proposed model of AI Architecture implementation within the Automotive
industry

(Source: Kotsiopoulos et al., 2021)

Instance-based Learning

A further supervised machine learning algorithm that can be implemented is "Instance-Based

Learning (IBL)", also known as Memory-Based Reasoning (MBR). It is possible to achieve high
accuracy of classification performance as well as solid performance using these algorithms,
which are used in both regression and classification applications. Nevertheless, if a large-scale
training dataset is required to be processed, there will be a significant increase in computational
complexity.

7
K Nearest Neighbours

The K-nearest neighbours algorithm is a member of the IBL heuristic family and can be utilized
to predict and assess machine configuration data, such as machine configuration and architecture,
motor performance, range, and cost (Kelley, 2020). It is also known as the K-nearest neighbours
algorithm. Also possible is the incorporation of this technology into the prediction of the
physical quality of intermediate products in interconnected manufacturing processes, as well as
the manufacturing costs associated with automobile engine components.

Logistic Regression/ Multiple Regressions

Multiple Regressions is a statistical technique that expresses how only one dependent variable Y
is dependent linearly on a handful of predictive variables X. The Multiple Regression method
with n predictor variables X1, X2... Xn and a response variable Y can be expressed as follows:

y = b0 + b1x1 + b2x2 + …. + bnxn + e (2.)

The model's residual terms, e, are represented by the variables, and the distribution presumption
is grounded upon the application of the residuals representation by A later inference on the
remaining model parameters will be possible thanks to term (e) The regression coefficients are
represented by the numbers b0, b1, b2,...bn. Furthermore, Logistic Regression is a slightly
different approach (generalized linear model) from Linear Regression in that it takes a longer
time to complete. The equation that adequately describes how it operates is as follows:

y = 1 whenever, b0+b1x1 +e > 0 else y =0.

In the case of the proposed report under consideration, the predictive analysis of the Australian
Automotive industry has been considered by evaluating multiple linear regressions within
Microsoft Excel.

8
Methodology

Predictive Analysis for modeling sample data- Automotive Fuel Organizations in

Australia
The Automotive fuel price data of Australia that has been considered for Predictive data analysis
has been tabulated as follows

Date Automotive fuel, Australia MQ20 Automotive fuel price level

Jun-15 95.1
Sep-15 93.5
Dec-15 88.2
Mar-16 79.4
Jun-16 84.1
Sep-16 81.7
Dec-16 87.2
Mar-17 92.2
Jun-17 89.9
Sep-17 87.8
Dec-17 96.9
Mar-18 97.8
Jun-18 104.5
Sep-18 106
Dec-18 103.4
Mar-19 94.4
Jun-19 104
Sep-19 101.9
Dec-19 106.4
Mar-20 100 100
Jun-20 80.7 100
Sep-20 88.3 100
Dec-20 88.7 100

9
Mar-21 96.4 100
Jun-21 102.7 100

Table 1: Automotive fuel price level comparison (March 2020 = 100.0)

(Source: Australian Bureau of Statistics, Consumer Price Index, Australia June 2021)

The forecast sheet analysis on the basis of automotive fuel consumption corresponding to
monthly data has been depicted below

Figure 2: Automotive Fuel consumption of Australia vs Month

(Source: Self-developed in MS Excel)

10
 Figure 3: Automotive Fuel consumption of Australia vs Month

(Source: Self-developed in MS Excel)

The Regression analysis table has been developed where the dependent variable under
consideration is the month and the independent variable under consideration is the automotive
fuel consumption data of Australia. The Regression table developed in Microsoft Excel is
described as given below

Regression Statistics
Multiple R 0.390637284
R Square 0.152597487
Adjusted R Square 0.1157539
Standard Error 6.920741561
Observations 25

11
 ANOVA
df SS MS F Significance F
198.37 4.1417
Regression 1 198.3767337 7 7 0.053520316
47.896
Residual 23 1101.623266 7
Total 24 1300

P-
Coefficie Standard valu Lower Upper Lower Upper
nt--ents Error t Stat e 95% 95% 95.0% 95.0%
- - - -
19.719984 16.13703 1.22 0.234 53.1019 13.662 53.1019 13.6620
Intercept 3 982 203 072 9454 02594 9454 2594
- -
Automotive 0.3479072 0.170950 2.03 0.053 0.00573 0.7015 0.00573 0.70154
fuel, Australia 85 662 5133 52 1103 45672 1103 5672

Table 2: Regression coefficients/statistics of Automotive fuel consumption vs. Month

(Source: Self-created in MS Excel)

Organizational Benefits
In the contexts pertaining to automotive fuel industry operations within Australia, it has to be
emphasized that the future statistics based on the forecast analysis corresponding to automotive
fuel consumption within Australia has been ascertained (Kotsiopoulos et al., 2021). Moreover,
the regression statistical analysis has been carried out wherein the dependent variable under
consideration is the month and the independent variable under consideration is the monthly
automotive fuel consumption figures. The R Square analysis that has been carried out outlines
the fact that the automotive fuel usage and monthly consumption has a 15.2% variation across
the various months of fuel consumption with a standard error (SE=6.92704). Moreover, the

12
regression analysis reveals the fact that there is a negative intercept relationship between the
automotive fuel consumption data and monthly consumption data. Moreover, the regression
analysis reveals a positive cor-relationship between automotive fuel consumption and monthly
figures with (p <0.05). This is indicative of the fact that with the prevailing Covid 19 trend there
will be a consequent reduction in the values of automotive fuel consumption thereby leading to
decreased profitability in the case of the automotive fuel industry within Australia.

The forecast analysis has been carried out to reveal the estimated automotive fuel consumption
over the course of the upcoming months within Australia. In this context, the automotive fuel
consumption with due consideration of the lower and upper confidence bounds has been pre-
determined (Zou et al., 2020). The analysis carried out has revealed the fact that the monthly fuel
consumption is likely to remain confined within the range of 100-110 in respect to the ongoing
Covid 19 pandemic situation. The lower confidence bound indicates the reduction in the
automotive fuel price value in the case of Australia with values dipping down from 61.2 to 53.7.
This is indicative of decreasing profit margins within the automotive fuel industry within
Australia. On the other hand, upper confidence bound estimation is indicative of the rise in
automotive fuel price levels. This is indicative of increasing profit margins within the automotive
fuel industry of Australia.

Conclusion
In order to conclude, it has to be mentioned that the following report has been prepared to outline
the utility of Big data and predictive analytics techniques in the case of the Automotive fuel
industry within Australia. It also underlines the benefits associated with effective AI/ML
applications in automobile industry operations. The application potential associated with
Machine learning (ML) and Artificial Intelligence (AI) within the automotive fuel organization
operations has been emphasized. Moreover, the Statistical Learning Theory (SLT) an industry-
standard supervised machine learning framework has been recommended. The predictive
analysis based on regression analysis of monthly fuel automobile fuel consumption and
automobile price levels has been suggested. The regression analysis has revealed a negative
intercept relationship between the automotive fuel consumption data and monthly consumption
data. Moreover, a positive correlationship between automotive fuel price levels and monthly

13
consumption has been ascertained via the regression analysis. Furthermore, the forecast analysis
has been conducted with the estimation of the expected profitability margins or revenues and
return on investment (ROIs) over the course of the upcoming months within the Australian
automobile sector.

14
Reference List
Journals

Abduljabbar, R., Dia, H., Liyanage, S. and Bagloee, S.A., 2019. Applications of artificial
intelligence in transport: An overview. Sustainability, 11(1), p.189.

Agrawal, A., Gans, J. and Goldfarb, A., 2019. Economic policy for artificial
intelligence. Innovation Policy and the Economy, 19(1), pp.139-159.

Attaran, M. and Deb, P., 2018. Machine learning: the new'big thing'for competitive
advantage. International Journal of Knowledge Engineering and Data Mining, 5(4), pp.277-305.

Attaran, M. and Deb, P., 2018. Machine learning: the new'big thing'for competitive
advantage. International Journal of Knowledge Engineering and Data Mining, 5(4), pp.277-305.

Balashanmugam, N., 2021. Perspectives on additive manufacturing in Industry 4.0. In Additive

Manufacturing (pp. 127-150). Woodhead Publishing.

Dagnaw, G., 2020. Artificial intelligence towards future industrial opportunities and challenges.

Davenport, T.H., 2018. From analytics to artificial intelligence. Journal of Business

Analytics, 1(2), pp.73-80.

Dogru, A.K. and Keskin, B.B., 2020. AI in operations management: applications, challenges and
opportunities. Journal of Data, Information and Management, 2(2), pp.67-74.

Hyder, Z., Siau, K. and Nah, F.F.H., 2018. Use of Artificial Intelligence, Machine Learning, and
Autonomous Technologies in the Mining Industry. MWAIS 2018 Proceedings, 43.

Johnston, I., 2021. Artificial Intelligence and the Ethics Behind It.

Kelley, J.N., 2020. IMPACT OF ARTIFICIAL INTELLIGENCE, MACHINE LEARNING,

AND AUTOMATION IN OPERATIONS MANAGEMENT: AN ANALYSIS OF
HEALTHCARE, MANUFACTURING, AND RETAIL SECTORS.

15
Kotsiopoulos, T., Sarigiannidis, P., Ioannidis, D. and Tzovaras, D., 2021. Machine Learning and
Deep Learning in Smart Manufacturing: The Smart Grid Paradigm. Computer Science
Review, 40, p.100341.

Wang, W. and Siau, K., 2019. Artificial intelligence, machine learning, automation, robotics,
future of work and future of humanity: A review and research agenda. Journal of Database
Management (JDM), 30(1), pp.61-79.

Zohuri, B. and Rahmani, F.M., 2019. Artificial Intelligence Driven Resiliency with Machine
Learning and Deep Learning Components. International Journal of Nanotechnology &
Nanomedicine, 4(2), pp.1-8.

Zou, H., Chen, G., Xie, P., Chen, S., He, Y., Huang, H., Nie, Z., Zhang, H., Bala, T., Tulip, K.
and Wang, Y., 2020. Validate and Enable Machine Learning in Industrial AI. arXiv preprint
arXiv:2012.09610.

16