2022 International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT)

Prediction of Defective Products Using Logistic

Regression Algorithm against Linear Regression
Algorithm for Better Accuracy
2022 International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT) | 978-1-6654-5193-2/22/$31.00 ©2022 IEEE | DOI: 10.1109/3ICT56508.2022.9990653

Vasantha Naga Vasu Surendran. R Saravanan. M.S

Department of Computer Science and Department of Computer Science and Department of Computer Science and
Engineering, Engineering, Engineering,
Saveetha School of Engineering, Saveetha School of Engineering, Saveetha School of Engineering,
Saveetha Institute of Medical and Saveetha Institute of Medical and Saveetha Institute of Medical and
Technical Sciences, Technical Sciences, Technical Sciences,
Chennai, India. Chennai, India. Chennai, India.
[email protected] [email protected] [email protected]

Madhusundar. N
Department of Computer Science and
Engineering
Saveetha Institute of Medical and
Technical Sciences,
Chennai, India.
[email protected]

Abstract—The aim of the research investigation is being IEEE Explorer published 78 research papers, and Google
done to increase the precision and prediction of defective online Scholar published 65 articles. The most findings of the
products using management to find the defective online proposed research are that an optimal online defective product
products using a machine, A novel logistic regression classifier strategy is necessary to achieve the target control system [3].
is applied to the online products dataset that consists of 3197 The illustration of the entire purely fictitious concave in profit,
records with 10 attributes and sample size=102. A framework recovery time, unit cost, and final stock level. A stochastic
for finding defective products in new defective online products inventory model should be designed for defective products
logistics comparing novel logistic regression and linear including shortage and expiry dates are proposed [4]. This
regression algorithms has been developed. The accuracy and
model is enlarged on the premise that production is routinely
precision of the models were evaluated and recorded. The linear
regression algorithm gives 76.0% accuracy in predicting the
paused for maintenance to prevent breakdowns and
defective rate, whereas the novel logistic regression algorithm interruptions in defective online products [5]. The signed
predicts the same with 84.2% accuracy. There exists a technique of distance is used to make the approximate cost of
statistically different in a way that both (p=0.03; p0.05) Linear the defective online products clear and its methodology is
regression and logistic regression). The performance of novel proposed to reduce the expenses involved [6]. By using an
logistic regression is significantly better and gives more iterative algorithm, the result accuracy was 73% [7]. The
appropriate results than the linear regression in finding linear regression algorithm gave the result of 73%. The
defective rates for online products. random forest algorithm used an SCM dataset which gave an
accuracy of 79% [8]. A mathematical model is proposed
Keywords—Machine learning, Linear Regression, Novel which represents an integrated production of online defective
Logistic Regression, Defective Products, Logistics, Defective Rate. products strategy for food products in multi-buyer and single
seller online defective products [9]. The boosting algorithm
I. INTRODUCTION used car supply data which gave 74% of accuracy. Previous
In this competitive world of marketing, manufacturers are research projects our team has worked on span a variety of
forced to pick a suitable online sales management using a fields [10-16]. Due to the expanding trend in this industry, we
logistics algorithm to economize their online defective choose to investigate this project.
product system is the procedure for planning, implementing,
In the existing research, they designed a vendor buyer
and monitoring the flow of information of goods in the online
inventory model for finding defective online products. It
defective products. It is important for reducing general
includes a quality check at the buyers' end [17]. The
expenses by reducing production and purchasing costs and
performance of linear regression models is lesser than other
enhancing the customer services like delivery of the correct
classification algorithms. However, it is critical to remember
product in the proper quantity at the proper time. To achieve
the following: analyze and compare different classification
customer satisfaction [1]. In day-to-day life, online defective
algorithms to see which ones are the most accurate. As a
products management is enhancing its services to obtain
result, the project's goals are as follows: at comparing the
success in the business of online defective products and
accuracy of novel logistic regression and linear regression
logistics [2]. It provides real-time analytical online defective
algorithms in finding the defective products. In Logistic
product systems to control the flow of goods in the defective
Regression Algorithm-Based Product Recommendation
online products. It also has computerized shipping and
System Model [18] discussed about the accuracy only our
tracking of goods.
proposed system discussed about accuracy and precision for
Recently, several kinds of network traffic detection linear regression and logistic regression algorithm.
using machine learning have been the subject of research.

Authorized licensed use limited to: COVENTRY UNIVERSITY. Downloaded on January 02,2024 at 05:06:12 UTC from IEEE Xplore. Restrictions apply.
978-1-6654-5193-2/22/$31.00 ©2022 IEEE 161
 2022 International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT)

II. PROPOSED MATERIALS AND METHODS

This study was conducted at saveetha school of a0=ym-a1*xm
Engineering's Machine Learning Laboratory. Linear
regression and the innovative logistic regression algorithm are DOFOR i = 1,n
used in this study. A power analysis was performed and st = st +(yi = ym)2
utilizing the g-power software. The G method was used to
estimate the number of samples for each category. g-power Sr = sr +(yi - a1*xi - a0)2
tool and resulted in 102 samples [19] having a power value of END DO
80% of g, Each group has 102 samples. In implementation part
logistic regression and linear regression algorithm certainly syx = (sr/(n-2))0.5
spending python. It’s possible that if we run the algorithm 102 r2=(st - sr)/st
times, we won't get different results. As a result, we are
limiting machine learning algorithms by altering the training End process
data and testing data percent for five times instead of 102
times. B. Novel Logistic regression algorithm
A related sort of supervised learning approach is novel
The SCMS_Delivary_History.csv [20], dataset utilized in logistic regression. In machine learning that deals with
this study, was obtained from the Kaggle website and contains classification problems and produces a distinct result. A
11000 rows and has 10 attributes, some of the attributes used produces a distinct result. It predicts the probability of a target
were product id, Line item value, quantity, weight, freight variable. The output is transferred to return a probability
cost, etc. The information is gathered to forecast the outcome. value, we use a logistic sigmoid function. It’s a probabilistic
Defective rate of online defective products before shipping. analytic algorithm that predicts outcomes. It makes use of a
Generally, to split the training and testing sets, an 80:20 ratio complicated cost function called the logistic function. The
is used. The data model is built with the help of a novel logistic sigmoid function is used to convert the predictions to
regression algorithm applied to training data, based on that the probability. Datasets are of the form {(x1, 1), (x2, 0), ..}. In
test data and result accuracy are predicted. this case, in this case, the linear regression will fail miserably.
A. Linear regression algorithm Classifying examples as positive (yi=1), or negative (yi=0).
A supervised machine learning method is a linear Pseudo code of Logistic Regression algorithm
regression technique classification algorithm that gives
Input: Supply chain Training dataset
continuous output. A linear regression algorithm is a
supervised machine learning algorithm that gives the output in Output: Accuracy
the form of a line. The relationship between one or more
features (Independent Variables) and a continuous variable Steps:
(Dependent variable) is shown. It deals with the regression Initialize all parameter
problems and gives continuous output. There are many types
of linear regression. If we use one feature for processing, then B0, B1,- - - - - - -etc
it is known as Univariate Linear regression. Likewise, Predict dependent Variable h(xi)
multiple linear regression if we use multiple features. Apart
from different types, a simple linear regression model was Calculate Cost Function
chosen. Cost (h(0)(X),y)
Pseudo code for Linear Regression algorithm Calculate gradient for cost function
Input: Supply chain dataset Update all parameters
Output: Accuracy Repeat calculating prediction value and cost function
SUB: Reg(x, y, a1, a0, n, syx) End process
Sumx = sumxy = st = 0 Therefore we are using a novel logistic regression
Sumy = sumx2 = sr = 0 algorithm. The Anaconda and Spyder C have been used to test
the suggested model. Computer used is with an 8GB RAM, I5
DOFOR 1 = 1,n CPU, 256GB HDD. In statistical Analysis independent
Sumx = sumx +xi variables are product Id, quantity, and purchased and the
accuracy of finding the defective products is the variable that
Sumy = sumy +yi is dependent (Defect percent) used for analysis. Finally, there
Sumxy = sumxy+xi* xi are the accuracies obtained from the two groups are processed
in SPSS is social science statistical software that may be used
Sum2= sumx2 +xi* xi to analyse data .Statistical analysis of both algorithms.
END DO
Xm = sumx/n
ym = sumy/n
a1 = (n*sumxy-sumx*sumy)/ (n*sum2-sumx*sumx)

Authorized licensed use limited to: COVENTRY UNIVERSITY. Downloaded on January 02,2024 at 05:06:12 UTC from IEEE Xplore. Restrictions apply.
162
 2022 International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT)

TABLE I. DURING THE EVALUATION, ACCURACY AND PRECISION The accuracy of these two algorithms is not dependent on
the data size. Statistical analysis of the results is obtained
Trainin Test Set Accurac Accurac Precisio Precisio
using SPSS in Table II. This analysis revealed that the mean
g Set (%) y of y of n of n of of the precision of Novel term "logistic regression" refers to a
(%) Linear logistic Linear logistic method of 85.82% and that of linear regression is 74.54%.
regressi regressi regressi regressi Similarly, the standard deviation of Novel Logistic Regression
on on on on is 2.217, and that of linear Regression is 4.320. The standard
error mean of Novel Logistic Regression is 1.108 and is less
90 10 76.0 84.2 76.3 89.3
than the standard error means of linear regression which is
2.160. A t-test using an independent sample was conducted to
80 20 75.2 83.1 75.5 88.6
compare the two groups, The statistical significance of the
70 30 74.6 82.7 74.2 87.9
result analysis is depicted 3 in Table P has a computed value
of 0.039, and 0.05 for the alpha.
60 40 73.8 81.1 74.0 87.1
TABLE II. STATISTICALLY SIGNIFICANT DIFFERENCE IN PRECISION
AND ACCURACY BETWEEN THE ALGORITHMS
50 50 73.1 80.3 73.6 86.5

Group N Mean Deviation Error Mean

from the (Standard
III. EXPRIMENTAL RESULTS norm Error)
Data pre-processing is done followed by data
Linear 5 74.54 2.217 1.108
normalization, which scales the data. Data cleaning is the next accuracy 5 85.82 4.320 2.160
step, where unwanted data is removing. Then the information logistic
is separated into training and testing data. Training data with
80%on the entire dataset, data validation with 20%on the Precision 5 74.72 2.217 1.108
entire dataset, this will be examined. Using a test set linear 5 87.88 5.231 2.236
repeatedly by altering the percentage of testing and testing logistic
data, Table I shows the results of the test for 5 trials. Fig. 1
observed that there exists a statistical difference between both
linear regression and logistic regression. It shows (p=0.03;
p0.05) in the independent sample t-test. It shows (p=0.03;
p0.05) in the independent sample t-test. It shows (p=0.03;
p0.05) in the independent sample t-test. The experimental
findings support the findings of the current investigation.
Finds the defective products with an accuracy of Defective %
and a precision of 76.3%. In contrast, the proposed system
obtained 84.2% of accuracy and a precision of 89.3% for
finding defective products.
A line graph is plotted for the accuracy obtained by
performing 5 trials using linear regression and novel logistic
regression algorithms. The trial number is taken along the
accuracy is measured along the y-axis and the x-axis. Both line
graphs are not showing any uniform upward trend or Fig. 2. Comparing the Accuracy of Logistic Regression ( blue ) and Linear
downward trend. Each line graph in this figure has its own Regression (orange) throughout the trials (X-Axis: Trial No. and Y-Axis:
Accuracy)
trend; therefore, there is no direct relation between the trials
no.
The results correspond to identical variances because the
calculated value of 'p' is less than the value of alpha in the
Table III. Assumed are taken into account for research, when
the amount of "t" is negative, implying that the mean is
negative Of Novel Logistic Regression is statistically superior
to the average of linear regression. The confidence interval for
the mean difference is [-21.217, 8.808] and 0 falls within this
interval, hence the results are statistically significant at alpha
= 0.05. The Novel Logistic regression model obtained 84.2%
accuracy, and linear regression obtained an accuracy of 76.0%
is shown in Fig. 2. The suggested approach outperforms other
algorithms in terms of performance. The algorithm
outperforms them all. Novel Logistic regression obtained
greater results than the linear regression algorithm in Fig. 3.

Fig. 1. Evaluation, accuracy and precision

Authorized licensed use limited to: COVENTRY UNIVERSITY. Downloaded on January 02,2024 at 05:06:12 UTC from IEEE Xplore. Restrictions apply.
163
 2022 International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT)

TABLE III. COMPARISON OF THE SIGNIFICANCE LEVEL AT P=1 FOR LINEAR REGRESSION AND LOGISTIC REGRESSION TECHNIQUES 0.039

Parameters F sig. t df sig.(2-tailed) Difference in Difference Lower Upper

the mean in standard confidence confidence
error by 95% level of
95%

Precision .833 .039 6.075 6 <.001 14.750 2.428 8.80857 20.691

6.075 4.478 .003 14.750 2.428 8.28288 21.217

Accuracy .833 .039 -6.075 6 <.001 -14.750 2.428 -20.691 8.808

-6.075 4.478 .003 -14.750 2.428 -21.217 8.282

Fig. 3. Axis X: Logistic Regression vs. Linear regression Algorithm Y Axis: Mean detection accuracy

IV. DISCUSSION
The proposed approach had a marginal accuracy gain of recovery time, unit price, and ending stock level the profit
8.2 percent. It is also worth noting that the Novel Logistic is strictly pseudo-concave. Here, the quality check is done at
regression algorithm in classifying the defective rate achieves the customer end [28]. The practices of mass production
greater precision and specificity. The results are statistically industries were identified and based on that decisions on
significant at 0.05 p-value, The improvement in accuracy of defective products have been taken into consideration [29].
the novel logistic regression algorithm is due to the better In similar findings, they performed quality checks in the
classification of defect product rate. It gives the Area under initial stages, i.e., the product is monitored in every step in the
Curve (AUC) of 0.82 which shows that the performance of the production process so that if any error is found, they will
algorithm on the data collected is better. Although the class correct it instantly. Therefore, cost reduction and quality
distribution in the dataset could be a factor in improving the improvement are achieved. In the opposing findings, they
accuracy, it is not analysed in this study. An optimal online have performed the quality check at the customer end which
defective product strategy is necessary to achieve the target leads to reverse reworking of the product. It affects the cost
control system [21]. A stochastic inventory model should be and time of the manufacturer and brings them losses.
designed for defective products including shortage and expiry Therefore, this study is more effective than the one that
date is proposed [22-23]. preceded it.
This model is enlarged by assuming that, to prevent Our university has been successful in a number of areas
failures and disruptions, production is regularly halted for and is committed to undertaking high-quality, evidence-based
maintenance of defective online products [24]. The research [30]. We hope that this study contributes to this
manufacturer checks the quality of the product in the initial illustrious history, owing to some uncertainties, the fuzzy
stages, to achieve the quality target which satisfies the defective rate has little disturbance. Online defective product
customer [25-27]. The illustration of the entire in terms of management still has some errors in finding defective rates

Authorized licensed use limited to: COVENTRY UNIVERSITY. Downloaded on January 02,2024 at 05:06:12 UTC from IEEE Xplore. Restrictions apply.
164
 2022 International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT)

and poor quality checking processes. We improved the [9] K. Haribabu, S. Muthukrishnan, S. Thanikodi, G. A. Antony, and V.
prediction of in this research, defective rate to save the Venkatraman, “Investigation of Air Conditioning Temperature
Variation by Modifying the Structure of Passenger Car Using
production cost, lead time, and quality improvement. Computational Fluid Dynamics,” Thermal Science, vol. 24, no.1, pp.
495–98, 2020.
The system can be improved in the future to forecast the
phase of the defective items where the actual mistake was [10] M. Sureshbabu, Nivedhitha, K. Selvarasu, J. Kumar V, M.
Nandakumar, and D. Selvam, “Concentrated Growth Factors as an
made. On datasets models are trained and tested using a Ingenious Biomaterial in Regeneration of Bony Defects after Periapical
normal standardizer in this work. Other algorithms can be Surgery: A Report of Two Cases,” Case Reports in Dentistry, no.
used in the same way, but this one is more successful. The 7046203, 2019.
number of samples directly influences the model's accuracy [11] M. G. Mathew, S. R. Samuel, A. J. Soni, and K. B. Roopa, “Evaluation
trained; if there are more samples, we will get appropriate of adhesion of Streptococcus mutans, plaque accumulation on zirconia
results. and stainless steel crowns, and surrounding gingival inflammation in
primary molars: randomized controlled trial,” Clinical Oral
Investigations, vol. 24, no. 9, pp. 3275–3280, 2020.
V. CONCLUSION
[12] R. Uthayakumar, and S. Hemapriya, “Integrated Vendor-Buyer
Within the parameters of this research, the Novel Logistic Inventory Model in a Supply Chain with Variable Lead Time and Setup
regression algorithm gives more significant results than the Cost Reduction,” International Journal of Process Management and
Algorithm for linear regression from an accuracy perspective. Benchmarking, vol. 9, no. 3, pp. 323–350 2019.
The novel Logistic regression Algorithm (84.2%) achieved an [13] Muriana, Cinzia, “An EOQ Model for Perishable Products with Fixed
Shelf Life under Stochastic Demand Conditions,” European Journal of
improvement of 4.2% of accuracy Over the Linear regression Operational Research, vol. 255, no. 2, pp. 388–396, 2016.
algorithm (76.0%). The factors that contributed to this [14] M. Sivaprakash, H. Krishnaswamy, S. Thanikodi, D. Sundaresan, and
improvement in accuracy are the type of method used, the V. Venkatraman, “Support Vector Machine for Modelling and
principle used for predicting SCM for finding defective Simulation of Heat Exchangers,” Thermal Science, vol. 24, no. 1, pp.
products, and the algorithm chosen for classification. As a 499–503, 2020.
result, it can be argued that Novel Logistic Regression, is the [15] J. Pc, T. Marimuthu, and P. Devadoss, “Prevalence and Measurement
outcome of the classifier, gives adequate precision and of Anterior Loop of the Mandibular Canal Using CBCT: A Cross
Sectional Study,” Clinical Implant Dentistry and Related Research,
accuracy to Linear Regression. 2018.
ACKNOWLEDGMENTS [16] S. Priyan and P. Mala. “Optimal Inventory System for Pharmaceutical
Products Incorporating Quality Degradation with Expiration Date: A
The authors would like to express their gratitude to Game Theory Approach,” Operations Research for Health Care, 2020.
Saveetha School of Engineering, Saveetha Institute of [17] T. Tamilvizhi, R. Surendran, C. A. T. Romero, M. Sadish Sendil,
Medical and Technical sciences (Formerly known as “Privacy Preserving Reliable Data Transmission in Cluster Based
Vehicular Adhoc Networks,” Intelligent Automation & Soft
Saveetha University) for providing the necessary Computing, vol. 34, no. 2, pp. 1265-1279, 2022.
infrastructure to carry out this work successfully. [18] K. S. Riya, R. Surendran, C. A. T. Romero, M. Sadish Sendil,
“Encryption with User Authentication Model for Internet of Medical
REFERENCES Things Environment,” Intelligent Automation & Soft Computing, vol.
[1] Alamerew, A. Yohannes, and D. Brissaud, “Circular Economy 35, no. 1, pp. 507–520, 2023.
Assessment Tool for End of Life Product Recovery Strategies,” Journal [19] A. Yeong-Hwi, P. Koo-Rack, K. Dong-Hyun, and C. Han-Jin,
of Remanufacturing, vol. 9, pp. 169–185, 2019. “Logistic Regression Algorithm-Based Product Recommendation
[2] E. Samira, S. Hosseini-Motlagh, and M. Nematollahi, “Proposing a System Model,” Journal of Computational and Theoretical
Delay in Payment Contract for Coordinating a Two-Echelon Periodic Nanoscience, vol. 18, no. 5, pp. 1429–1435, 2021.
Review Supply Chain with Stochastic Promotional Effort Dependent [20] https://www.kaggle.com/datasets/divyeshardeshana/supply-chain-
Demand,” International Journal of Machine Learning and Cybernetics, shipment-pricing-data
vol. 10, pp. 1037–1050, 2019. [21] G. K. J, G. S, S. Rajendran, J. S. Vimali, J. Jabez, and S. Srininvasulu,
[3] E. Devaraj, V. S. Apoorva, and N. A. Vardhan, “Syzygium Cumini “Identification of Cyber Threats and Parsing of Data,” 2021 5th
Extract Induced Reactive Oxygen Species-Mediated Apoptosis in International Conference on Trends in Electronics and Informatics
Human Oral Squamous Carcinoma Cells,” Journal of Oral Pathology (ICOEI), pp. 556-564, 2021.
& Medicine: Official Publication of the International Association of [22] D. A. Reda, I. H. Challoob and S. H. Omran, “The Use of Time Series
Oral Pathologists and the American Academy of Oral Pathology, vol. to Predict Defective Percentages is an Applied Study in The Ishtar
48, no. 2, pp. 115–21, 2019. Fireplace Laboratory,” 2020 2nd Al-Noor International Conference for
[4] F. Gusti, Y. Amer, S. Lee, and H. Prasetyo, “An Integrated Single- Science and Technology (NICST), pp. 13-17, 2020.
Vendor Multi-Buyer Production-Inventory Policy for Food Products [23] Samuel, S. Melvin, J. Bhattacharya, S. Raj, N. Santhanam, H. Singh,
Incorporating Quality Degradation,” International Journal of and N. D. P. Singh. “Efficient Removal of Chromium(VI) from
Production Economics, vol. 182, pp. 409–417, 2016. Aqueous Solution Using Chitosan Grafted Graphene Oxide (CS-GO)
[5] F. Lin, Y. Chan, and L. E. Cárdenas-Barrón, “Pricing and Lot-Sizing Nanocomposite,” International Journal of Biological Macromolecules,
Polices for Perishable Goods When the Demand Depends on Selling vol. 121, pp. 285–92, 2019 .
Price, Displayed Stocks, and Expiration Date,” International Journal of [24] Samuel, S. Raj, S. Acharya, and J. Chandrasekar Rao, “School
Production Economics, vol. 185, pp. 11–20, 2017. Interventions-Based Prevention of Early-Childhood Caries among 3-5-
[6] S. Gheena, and D. Ezhilarasan, “Syringic Acid Triggers Reactive Year-Old Children from Very Low Socioeconomic Status: Two-Year
Oxygen Species-Mediated Cytotoxicity in HepG2 Cells,” Human & Randomized Trial,” Journal of Public Health Dentistry, vol. 80, no. 1,
Experimental Toxicology, vol. 38, no. 6, pp. 694–702, 2019. pp. 51–60, 2020.
[7] J. Jerry, Ajitha, and H. Subbaiyan. “Different Treatment Modalities [25] R. Surendran and T. Tamilvizhi, “Friendly Online Technology
Followed by Dental Practitioners for Ellis Class 2 Fracture – A Development Cloud Service for Bahraini Students based on E-
Questionnaire-Based Survey,” The Open Dentistry Journal, vol. 14, no. Advisor,” 2018 International Conference on Innovation and
1, pp. 59–65, 2020. Intelligence for Informatics, Computing, and Technologies (3ICT), pp.
[8] K. Yang, M. S. A. Aboody, W. Alturaiki, S. A. Alsagaby, F. A. Alfaiz, 1–7, 2018.
V. Priya, Veeraraghavan, and S. Mickymaray, “Photosynthesized Gold [26] Shabani, Sonia, Abolfazl Mirzazadeh, and Ehsan Sharifi, “A Two-
Nanoparticles from Catharanthus Roseus Induces Caspase-Mediated Warehouse Inventory Model with Fuzzy Deterioration Rate and Fuzzy
Apoptosis in Cervical Cancer Cells (HeLa),” Artificial Cells, Demand Rate under Conditionally Permissible Delay in Payment,”
Nanomedicine, and Biotechnology, no. 47, no. 1, pp. 1938–46, 2019. Journal of Industrial and Production Engineering, vol. 33, no. 2, 2016.

Authorized licensed use limited to: COVENTRY UNIVERSITY. Downloaded on January 02,2024 at 05:06:12 UTC from IEEE Xplore. Restrictions apply.
165
 2022 International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT)

[27] Sharma, Parvarish, M. Mehta, D. S. Dhanjal, S. Kaur, G. Gupta, H.

Singh, and L. Thangavelu, “Emerging Trends in the Novel Drug
Delivery Approaches for the Treatment of Lung Cancer,” Chemico-
Biological Interactions, vol. 309, no. 108720, 2019.
[28] Soares, J. Cláudio, S. Sousa, and A. Tereso, “Industry Practices on the
Rework of Defective Products: Survey Results,” The TQM Journal,
vol. 32, no. 6, pp. 1177–1196, 2020.
[29] S. Amanda, C. N. Rosyidi, and E. Pujiyanto, “Product Quality
Improvement Model Considering Quality Investment in Rework
Policies and Supply Chain Profit Sharing,” Journal of Industrial
Engineering International, vol. 15, pp. 637–649, 2019.
[30] T. Tamilvizhi, R. Surendran, R. M. Bommi, “Radio Frequency
Identification (RFID) based ubiquitous health care data handling,” IOP
Conference Series: Materials Science and Engineering, vol. 994, no.
1, 2020.

Authorized licensed use limited to: COVENTRY UNIVERSITY. Downloaded on January 02,2024 at 05:06:12 UTC from IEEE Xplore. Restrictions apply.
166