Improving Big Data Intelligence Using Entropy Weighted Method for Cloud-Based AutoML Evaluation

Authors

• KIRAN KUMAR MANDULA SAMUEL Osmania University

Abstract

We  explore  the topic of trust in the emerging field of automated machine learning (AutoML).  AutoML uses artificial intelligence methods to automate the creation and improvement of machine learning models, including feature engineering, algorithm selection, and hyperparameter optimization. This integration of technologies is now widely referred to as Big Data Science. In addition to supporting sophisticated analytical tasks, cloud computing provides a practical and cost-effective solution for managing the storage and computational needs of Big Data. In this paper, we examine the key components of the software stack that allow Big Data Science to be delivered as an easily accessible service to data scientists.The research on the use of AutoML in the cloud, assessed by the Entropy Weighted Method, is significant for democratizing big data intelligence. By automating machine learning processes and using  Entropy Weighted Method  to rank cloud platforms, this research empowers organizations and individuals to make informed decisions. It ensures that even non-experts without deep technical knowledge can use advanced analytics. This work highlights the ability to make objective, data-driven decisions in selecting optimal AutoML solutions, supporting businesses and researchers in maximizing data value, and improving competitiveness in a rapidly evolving digital landscape.Alternatives:LTE, UMTS, WiMAX, WLAN-I, WLAN-II, Evaluation parameter:Data transmission delay (PD),Packet delay variance (PJ), Data transfer rate (T), Packet loss rate (PL).The findings show that WiMAXholds its first position with the highest ranking, while LTE has the lowest ranking. Based on the Entropy Weighted Method, an analysis of the cloud that simplifies big data intelligenceshows that the WiMAXoccupies a leading position.

Keywords: AutoML, Cloud-Based Machine Learning, Data-Driven Decision Making,   Democratizing Data Science, AI in the Cloud          

Building on our discussion, we clarify these three key concepts.Access to information about how an AutoML system works and details about the models it generates is expected to foster trust through increased transparency. On the other hand, the absence of such information undermines trust in the system. As a result, we aim to highlight the most important types of information that an AutoML system should provide in its interface to build user trust. Existing AutoML systems already provide a variety of details about their operations and the models they produce. We generally categorize this information into three main categories: data-related details (such as how the data is transformed or preprocessed), engineering processes, and information about the produced models [1]. This paper presents findings from three studies aimed at understanding how the presence or absence of these types of information affects data scientists’ trust in an AutoML system. Our first study is development-oriented and involves semi-structured interviews to explore the full range of information included in a representative set of commercial AutoML products, as well as information that is commonly missing. We propose that adding this often-overlooked “hidden” information can improve the transparency of AutoML systems and, consequently, increase user trust. The second part of our study quantitatively assesses how the inclusion of this additional information, referred to as “transparency features,” affects trust ratings [2]. In addition, we provide a ranking of the relative importance of various information elements, or “nuggets,” in building trust in AutoML systems.These automation technologies typically handle only one aspect of the process, allowing data scientists to integrate these different parts to create a complete model development pipeline [4].

It is only in recent years that complete AutoML solutions have emerged that offer the ability to handle the entire data science workflow - data acquisition, model selection, deployment, and even continuous improvement.They found that data scientists are open to a collaborative future where humans and AI can work together to build models [5]. Lee et al. have used the “mixed-initiative” literature, arguing that AutoML and human users “can work collaboratively to achieve user goals.” This frames AutoML as a human-in-the-loop system.These new empirical studies expand our understanding of how data scientists perceive, interact with, and collaborate with AutoML tools. They also provide valuable design insights for improving the usability of AutoML systems. However, as highlighted by recent work, transparency and trust remain significant barriers to the widespread adoption of these tools. In the following subsection, we will delve deeper into the issues of transparency and trust in AutoML.AutoML is still a relatively new field, and only a few studies have specifically examined the issue of trust in AutoML systems. Some stages in the AutoML pipeline involve discrete search spaces (such as different algorithms in the model selection phase), while others rely on continuous search dimensions, making it challenging to visualize this information.We are in the midst of the era of Big Data [7]. Rapid advances in data storage, computing capabilities, digital devices, and networking infrastructure have created an ecosystem that is fueling the explosive growth of big data and the tools needed to create, manage, share, and analyze it. The term “big data” refers to the global rise of digital data generated through a variety of technologies and formats. Big data describes the result of continuous and significant advances in computing power, memory, storage, and data availability, which have fueled the development of new platforms designed to handle increasingly complex data science tasks. The relationship between big data and data science can be likened to that of crude oil and a refinery [8]. As a result, a multitude of tools for storing, processing, and analyzing big data have emerged that help transform complex information into meaningful patterns, insights, and knowledge. The sheer power of big data is reshaping every aspect of society.

The techniques and technologies behind big data science permeate every area of ​​business and research, impacting everything from the operations of modern organizations to the everyday decisions of digital citizens [10]. Big data analytics is driving change and progress in all sectors. Big data analytics is driving change and progress across all industries. In this article, we take a detailed look at the fundamental components of the software stack that enables Big Data Science as a service, accessible to data scientists "Data that requires scale, distribution, diversity, and/or timeliness to unlock new sources of business value" [12].Specifically, cloud computing represents a paradigm shift in how computing infrastructure is delivered. By centralizing resources in large-scale data centers, it reduces the costs and complexity of managing hardware and software. It provides users with access to unlimited computing resources, ensuring scalability by dynamically adding capacity as needed. It has revolutionized the IT industry by making computing more flexible and cost-effective through a pay-as-you-go model [13]. Cloud computing allows us to “think of computing as an application, with economies of scale that help reduce infrastructure costs.Major technology companies such as Amazon, Google, Microsoft, and IBM have invested heavily in building data centers and cloud services worldwide, which ensure reliability through redundancy in their supported infrastructure, platforms, and applications [15].The combination of big data and cloud computing has made it easier and more flexible than ever for anyone to engage in big data processing [16].

We emphasize analytics because the insights gained from interacting with big data ultimately provide societal benefits. Collectively, they foster productivity—a self-organizing capability that we believe is key to overcoming many of the big data challenges in this field. For example, NASA’s climate change data repositories alone were expected to grow to 350 petabytes by 2013, highlighting some of the major big data challenges facing climate science. The complexity of these data structures and their intricate interactions require specialized expertise for performance modeling in big data environments [18]. The term “big data” generally refers to scenarios involving enormous, often unstructured, and heterogeneous datasets that need to be analyzed, managed, and organized to support business objectives. The advent of big data has put significant strain on traditional data management approaches, as existing computational methods are inadequate to handle such vast and ever-expanding data volumes. This has fundamentally changed how modern data is processed [19]. In the IoT era, it is essential to conduct big data analytics within cloud data centers to deliver smart services such as intelligent healthcare solutions. Incorporating AI methods for data delivery in IoT environments involves the use of deep learning approaches that can effectively address uncertainties. In particular, deep learning offers high accuracy and robust performance, making it well suited for big data applications [20].

1. INTRODUCTION
2. Materials and Method

The advent of automated machine learning (AutoML) has revolutionized the way data-driven insights are generated, especially when combined with cloud computing. This synergy simplifies big data insights, making them accessible even to non-expert users. Together, they reduce technical barriers and help users transform large, complex datasets into actionable insights quickly and efficiently.A key challenge in this ecosystem is choosing the most appropriate AutoML tool or platform, given the vast number of options available.It uses a straightforward process that evaluates the performance of each alternative against the weighted criteria, normalizes the data, and calculates a comprehensive priority score.In applying the Entropy Weighted Method, the first step involves identifying relevant evaluation criteria and assigning appropriate weights that reflect their importance. Next, a decision matrix is ​​created that describes the performance of each alternative against these criteria (such as cloud-based AutoML platforms). The matrix is ​​then normalized to allow for direct comparison regardless of the different criteria. These scores are used to calculate the relative importance (Qi) and utility degree (Ui) and to determine the final ranking of the alternatives.The materials for this study include data on performance metrics (accuracy, cost, scalability, etc.) of various cloud-based AutoML platforms. The Entropy Weighted Method was used as an evaluation framework supported by tools such as Excel for data normalization and calculations.Materials:A set of AutoML tools available in cloud environments (e.g., Google AutoML, Amazon Sage Maker Autopilot, and Azure AutoML).A dataset of comparison criteria such as ease of use, scalability, cost, transparency, and performance metrics.Evaluation data obtained from vendor documentation, academic literature, and user surveys.

Alternatives: In modern wireless communications, several technologies play a key role in providing seamless connectivity and high-speed data access. LTE (Long Term Evolution) provides fast mobile internet with low latency, supporting applications from video streaming to IoT. UMTS (Universal Mobile Telecommunications System) builds on 3G capabilities, providing reliable voice and data services. WiMAX (Worldwide Interoperability for Microwave Access) is designed for high-speed broadband over long distances, useful in urban and rural areas. WLAN-I (Wireless Local Area Network, first generation) enables local wireless connectivity, commonly found in homes and offices. WLAN-II (second generation) further improves WLAN performance, supporting faster speeds and better reliability.

LTE: Provides higher data rates and mobility for modern smartphones and IoT.

UMTS: Provides voice and moderate data speeds over 3G networks.

WiMAX: Provides wireless broadband where fiber is not possible.

WLAN-I: Basic wireless LAN for home or small office use.

WLAN-II: Advanced wireless LAN with improved speed and coverage.

Evaluation parameter:

Data transmission delay (PD): The time it takes for a data packet to reach its destination from its source.

Packet delay variance (PJ): The fluctuations in the delay of packets that affect the consistency of data delivery.

Data transfer rate (T): The amount of data successfully moved across the network within a given time frame.

Packet loss rate (PL): The proportion of data packets that do not successfully deliver to their intended destination.

3. ANAZYE AND DISCUSSION

Table 1.The cloud that simplifies big data intelligence

Table 1 presents the performance of various wireless technologies in an AutoML environment in the cloud, aiming to simplify Big Data intelligence. Each technology is evaluated based on four key parameters: Data transmission delay (PD),Packet delay variance (PJ), Data transfer rate (T), Packet loss rate (PL). The data indicates that UMTS performs best in terms of packet delay (50.12 ms) and packet jitter (52.23 ms), providing reliable transmission for AutoML workloads. LTE achieves the highest throughput (5.5 Mbps), which is beneficial for data-intensive applications. WiMAX has moderate values, while WLAN-I and WLAN-II show high packet delays and low throughput, indicating potential limitations in demanding data science environments. Packet loss is relatively low across all networks, with UMTS and WiMAX performing particularly well (0.56% and 0.54%, respectively). These results highlight how different network technologies affect the real-time performance of cloud-based AutoML systems.

FIGURE 1.The cloud that simplifies big data intelligence

The bar graphs show the comparative performance of various wireless network technologies, LTE, UMTS, WiMAX, WLAN-I, and WLAN-II, based on four key metricsData transmission delay (PD),Packet delay variance (PJ), Data transfer rate (T), Packet loss rate (PL) .LTE and UMTS have low packet delays (60 ms and 50 ms, respectively) and jitter (55 ms and 52 ms), which are important for real-time data transfers in AutoML tasks. LTE also has high throughput (~5.5 Mbps), making it suitable for processing large amounts of data. WiMAX, WLAN-I, and WLAN-II exhibit high latency (above 80 ms) and low throughput (below 3 Mbps), indicating potential performance bottlenecks. All networks maintain relatively low packet loss, with UMTS and WiMAX performing particularly well.These diagrams illustrate how LTE and UMTS provide balanced throughput to support AutoML workloads in the cloud, even though the high latency in WLAN-based networks can impact responsiveness and performance.

Table 2. Normalized Data

The Normalized Data Table compares five wireless network technologies - LTE, UMTS, WiMAX, WLAN-I, and WLAN-II - on four key metrics: Data transmission delay (PD),Packet delay variance (PJ), Data transfer rate (T), Packet loss rate (PL).LTE stands out with the highest normalized throughput score (0.2970), making it the most suitable for high-data AutoML tasks, despite its moderate scores on latency and jitter. UMTS has low packet delay (0.1385) and jitter (0.2128), while maintaining reasonable throughput (0.2327), making it an efficient and balanced option for most applications. WiMAX, WLAN-I and WLAN-II show high latencies (above 0.22 for PD), which can slow down AutoML performance, although WLAN-II has the lowest jitter (0.1725), making it a good choice for less demanding tasks. Packet loss is relatively low across all networks, with UMTS (0.1534) and WiMAX (0.1479) excelling.

Table 3. This table provides equally assigned weights to four key network performance metrics Data transmission delay (PD),

This table provides equally assigned weights to four key network performance metrics Data transmission delay (PD),Packet delay variance (PJ), Data transfer rate (T), Packet loss rate (PL) . Each metric has a weight of 0.25 across various wireless technologies, including LTE, UMTS, WiMAX, WLAN-I, and WLAN-II. This even distribution indicates that all of these parameters are considered equally important for assessing the quality of service in these networks.Packet delay and jitter affect real-time communication quality, while throughput measures data transfer efficiency, and packet loss affects reliability. Assigning equal weights suggests a balanced approach where no single factor dominates the evaluation process. This reflects the holistic view that maintaining low latency and jitter, high throughput, and minimal packet loss collectively ensures optimal network performance. This method is very useful for fair comparison across different wireless standards, emphasizing that all four metrics are important in providing consistent and satisfactory user experiences.

Table 4 .Weighted Normalized Decision Matrix

The weighted normalized results matrix shows the performance scores of five wireless technologies - LTE, UMTS, WiMAX, WLAN-I, and WLAN-II - on four key metricsData transmission delay (PD),Packet delay variance (PJ), Data transfer rate (T), Packet loss rate (PL). These values ​​are obtained by normalizing the raw data and applying equal weights to each metric, reflecting their relative importance.LTE consistently shows high scores in throughput and packet loss, indicating strong performance in data transfer efficiency and reliability. UMTS scores are lower overall, especially in packet loss, indicating that it may have less reliable data transfer. WiMAX, WLAN-I, and WLAN-II have similar scores with small variations in packet delay and packet loss, showing competitive performance, but with some tradeoffs in latency and jitter.This matrix assists in objective comparison by balancing different metrics, enabling informed decisions when selecting the most appropriate technology based on network quality factors such as latency, jitter, throughput, and loss.

Table 5. Bi, Ci, Min (Ci)/Ci

The weighted normalized results matrix shows the performance scores of five wireless technologies - LTE, UMTS, WiMAX, WLAN-I, and WLAN-II - on four key metrics: Data transmission delay (PD),Packet delay variance (PJ), Data transfer rate (T), Packet loss rate (PL). These values ​​are obtained by normalizing the raw data and applying equal weights to each metric, reflecting their relative importance.LTE consistently shows high scores in throughput and packet loss, indicating strong performance in data transfer efficiency and reliability. UMTS scores are lower overall, especially in packet loss, indicating that it may have less reliable data transfer. WiMAX, WLAN-I, and WLAN-II have similar scores with small variations in packet delay and packet loss, showing competitive performance, but with some tradeoffs in latency and jitter. This matrix assists in objective comparison by balancing different metrics, enabling informed decisions when selecting the most appropriate technology based on network quality factors such as latency, jitter, throughput, and loss.

Table 6. Qi, Ui, Rank

This table shows the performance evaluation of five wireless technologies, LTE, UMTS, WiMAX, WLAN-I, and WLAN-II, using three metrics: Qi (overall performance index), Ui (usability value), and a final ranking based on these values. WiMAX emerges as the best performer with a Qi of 0.225 and a perfect usability score of 100, earning it the top ranking. WLAN-II follows closely with a Qi of 0.224 and a usability value of 99.5677, earning it the second spot.WLAN-I and UMTS show intermediate performance with usability values ​​of 89.3168 and 82.8518, ranking third and fourth, respectively. LTE has the lowest performance indices, with a Qi of 0.163 and a usability score of 72.3433, ranking fifth.This ranking indicates that WiMAX and WLAN-II offer the best performance, while LTE lags behind. Such assessments help network planners and engineers identify the most effective technology based on consistent performance and usability.

FIGURE 2. Rank

The bar charts provide comparative rankings and performance metrics – Qi, Ui, and Rank – of five wireless technologies: LTE, UMTS, WiMAX, WLAN-I, and WLAN-II. Each bar group represents the performance of the technologies on these metrics.From the charts, LTE consistently shows low values ​​across Qi, Ui, and Rank, highlighting its relatively weak performance. UMTS follows with moderate values, indicating better performance than LTE, but below its top competitors. WiMAX stands out with the highest Ui and Rank scores, underscoring its leading performance, closely followed by WLAN-II. WLAN-I also demonstrates strong performance, almost matching WiMAX and WLAN-II.These graphs underscore that WiMAX and WLAN-II are superior performers in terms of usage and rank, reflecting their superior network performance and suitability. On the other hand, LTE lags significantly behind, indicating the need for improvements or alternative options in high-performance scenarios.

4. Conclusion

The use of AutoML (automatic machine learning) in the cloud has significantly changed how organizations and individuals deal with big data intelligence. Entropy Weighted works by evaluating multiple criteria such as scalability, ease of use, cost-effectiveness, accuracy, and integration capabilities. Each criterion is weighted and scored for different AutoML solutions, allowing stakeholders to rank these platforms based on their relative performance. Platforms with higher accuracy and lower costs receive higher scores, ensuring an objective and data-driven selection process. This systematic approach not only simplifies the evaluation of AutoML options, but also helps align platform choices with business goals and technical requirements.Using the COPRAS methodology to evaluate cloud-based AutoML solutions provides a structured, transparent, and data-driven approach to selecting the most appropriate platform. As the demand for big data insights increases, AutoML in the cloud empowers organizations to gain insights quickly and efficiently, democratizing access to advanced analytics. Using COPRAS ensures that decision makers can objectively weigh key criteria, enabling better investments in AutoML platforms that best support their unique data-driven initiatives.

REFERENCE

1. Drozdal, Jaimie, Justin Weisz, Dakuo Wang, Gaurav Dass, Bingsheng Yao, Changruo Zhao, Michael Muller, Lin Ju, and Hui Su. "Trust in AutoML: exploring information needs for establishing trust in automated machine learning systems." In Proceedings of the 25th international conference on intelligent user interfaces, pp. 297-307. 2020.
2. Elshawi, Radwa, SherifSakr, Domenico Talia, and Paolo Trunfio. "Big data systems meet machine learning challenges: towards big data science as a service." Big data research 14 (2018): 1-11.
3. Schnase, John L., Daniel Q. Duffy, Glenn S. Tamkin, Denis Nadeau, John H. Thompson, Cristina M. Grieg, Mark A. McInerney, and William P. Webster. "MERRA analytic services: Meeting the big data challenges of climate science through cloud-enabled climate analytics-as-a-service." Computers, Environment and Urban Systems 61 (2017): 198-211.
4. Barbierato, Enrico, Marco Gribaudo, and Mauro Iacono. "Performance evaluation of NoSQL big-data applications using multi-formalism models." Future Generation Computer Systems 37 (2014): 345-353.
5. Sharma, Sugam. "Expanded cloud plumes hiding Big Data ecosystem." Future Generation Computer Systems 59 (2016): 63-92.
6. Al-Turjman, Fadi, HadiZahmatkesh, and Leonardo Mostarda. "Quantifying uncertainty in internet of medical things and big-data services using intelligence and deep learning." IEEE Access 7 (2019): 115749-115759.
7. Roh, Yuji, Geon Heo, and Steven EuijongWhang. "A survey on data collection for machine learning: a big data-ai integration perspective." IEEE Transactions on Knowledge and Data Engineering 33, no. 4 (2019): 1328-1347.
8. Barbierato, Enrico, Marco Gribaudo, and Mauro Iacono. "A Performance Modeling Language For Big Data Architectures." In ECMS, pp. 511-517. 2013.
9. Dai, Wenyun, LongfeiQiu, Ana Wu, and MeikangQiu. "Cloud infrastructure resource allocation for big data applications." IEEE Transactions on Big Data 4, no. 3 (2016): 313-324.
10. Zheng, Zhigao, Ping Wang, Jing Liu, and Shengli Sun. "Real-time big data processing framework: challenges and solutions." Applied Mathematics & Information Sciences 9, no. 6 (2015): 3169.
11. Shelestov, Andrii, MykolaLavreniuk, Vladimir Vasiliev, Leonid Shumilo, AndriiKolotii, BohdanYailymov, NataliiaKussul, and Hanna Yailymova. "Cloud approach to automated crop classification using Sentinel-1 imagery." IEEE Transactions on Big Data 6, no. 3 (2019): 572-582.
12. Xu, Yunbi, Xingping Zhang, Huihui Li, HongjianZheng, Jianan Zhang, Michael S. Olsen, Rajeev K. Varshney, Boddupalli M. Prasanna, and QianQian. "Smart breeding driven by big data, artificial intelligence, and integrated genomic-enviromic prediction." Molecular Plant 15, no. 11 (2022): 1664-1695.
13. Gill, Sukhpal Singh, ShreshthTuli, MinxianXu, Inderpreet Singh, Karan Vijay Singh, Dominic Lindsay, ShikharTuli et al. "Transformative effects of IoT, Blockchain and Artificial Intelligence on cloud computing: Evolution, vision, trends and open challenges." Internet of Things 8 (2019): 100118.
14. Kakulavaram, S. R. (2023). Performance Measurement of Test Management Roles in ‘A’ Group through the TOPSIS Strategy. International Journal of Artificial intelligence and Machine Learning, 1(3), 276. https://doi.org/10.55124/jaim.v1i3.276
15. Becker, Tilman, Edward Curry, AnjaJentzsch, and Walter Palmetshofer. "New Horizons for a Data-Driven Economy: Roadmaps and Action Plans for Technology, Businesses, Policy, and Society." (2016): 277-291.
16. Gupta, BrijBhooshan, Dharma P. Agrawal, Shingo Yamaguchi, and Michael Sheng. "Advances in applying soft computing techniques for big data and cloud computing." Soft Computing 22 (2018): 7679-7683.
17. Qolomany, Basheer, Ala Al-Fuqaha, Ajay Gupta, DrissBenhaddou, SafaaAlwajidi, JunaidQadir, and Alvis C. Fong. "Leveraging machine learning and big data for smart buildings: A comprehensive survey." IEEE access 7 (2019): 90316-90356.
18. Wang, Kun, Yun Shao, Lei Shu, Chunsheng Zhu, and Yan Zhang. "Mobile big data fault-tolerant processing for ehealth networks." IEEE Network 30, no. 1 (2016): 36-42.
19. Peram, S. R (2023). Advanced Network Traffic Visualization and Anomaly Detection Using PCA-MDS Integration and Histogram Gradient Boosting Regression. Journal of Artificial Intelligence and Machine Learning, 1(3), 281. https://doi.org/10.55124/jaim.v1i3.281
20. Wang, HuiZhe, Guo Wen Lin, Jian Qin Wang, Wan Lin Gao, Yi Fei Chen, and Qing Ling Duan. "Management of big data in the internet of things in agriculture based on cloud computing." applied mechanics and materials 548 (2014): 1438-1444.
21. Hao, Qian, and Lele Qin. "The design of intelligent transportation video processing system in big data environment." Ieee Access 8 (2020): 13769-13780.
22. Jiang, Jing, Jie Lu, Guangquan Zhang, and Guodong Long. "Optimal cloud resource auto-scaling for web applications." In 2013 13th IEEE/ACM international symposium on cluster, Cloud, and Grid Computing, pp. 58-65. IEEE, 2013.
23. Methods in Medicine, Computational and Mathematical. "Retracted: Influential Usage of Big Data and Artificial Intelligence in Healthcare." (2023): 9854236.
24. Sridhar Kakulavaram. (2024). Artificial Intelligence-Driven Frameworks for Enhanced Risk Management in Life Insurance. Journal of Computational Analysis and Applications (JoCAAA), 33(08), 4873–4897. Retrieved from https://www.eudoxuspress.com/index.php/pub/article/view/2996
25. Chen, Hong-Mei, Rick Kazman, Serge Haziyev, and OlhaHrytsay. "Big data system development: An embedded case study with a global outsourcing firm." In 2015 IEEE/ACM 1st International Workshop on Big Data Software Engineering, pp. 44-50. IEEE, 2015.
26. Roy, Jagannath, Haresh Kumar Sharma, SamarjitKar, EdmundasKazimierasZavadskas, and Jonas Saparauskas. "An extended COPRAS model for multi-criteria decision-making problems and its application in web-based hotel evaluation and selection." Economic research-Ekonomskaistraživanja 32, no. 1 (2019): 219-253.
27. Raghavendra Sunku. (2023). AI-Powered Data Warehouse: Revolutionizing Cloud Storage Performance through Machine Learning Optimization. International Journal of Artificial intelligence and Machine Learning, 1(3), 278. https://doi.org/10.55124/jaim.v1i3.278
28. Özdağoğlu, Aşkın, GülinZeynepÖztaş, Murat Kemal Keleş, and VolkanGenç. "An integrated PIPRECIA and COPRAs method under fuzzy environment: A case of truck tractor selection." Alphanumeric Journal 9, no. 2 (2021): 269-298.
29. Săraru, CodruțȘtefan. "Selection of the optimum artificial lift method, on the basis of aras, copras and topsis models." Petroleum-Gas University of Ploiesti Bulletin, Technical Series 68, no. 2 (2016): 28-40
30. Ballamudi, S. "Interleaved Feature Extraction Model Bridging Multiple Techniques for Enhanced Object Identification" Journal of Artificial Intelligence and Machine Learning., 2023, vol. 1, no. 2, pp. 1-7. doi: https://doi.org/10.55124/jaim.v1i2.253
31. Wang, Ze-Ling, Jian-Xin You, Hu-Chen Liu, and Song-Man Wu. "Failure mode and effect analysis using soft set theory and COPRAS method." International Journal of Computational Intelligence Systems 10, no. 1 (2017): 1002-1015.
32. Ridoy, MdMehediHasan, ShadmanSakib, and MdAriful Islam. "A Comprehensive Analysis of Mobile Phone Selection by Using MCDM Methods throughout AHP, COPRAS, and TOPSIS on Assessment of Customer Preferences." (2024).
33. Krishankumar, R., Harish Garg, KarthikArun, AbhijitSaha, K. S. Ravichandran, and SamarjitKar. "An integrated decision-making COPRAS approach to probabilistic hesitant fuzzy set information." Complex & Intelligent Systems 7, no. 5 (2021): 2281-2298.
34. Garg, R., and D. Jain. "Fuzzy multi-attribute decision making evaluation of e-learning websites using FAHP, COPRAS, VIKOR, WDBA." Decision Science Letters 6, no. 4 (2017): 351-364.
35. Adari, V. K. (2024). Interoperability and Data Modernization: Building a Connected Banking Ecosystem. International Journal of Computer Engineering and Technology, 15(6), 653–662.
36. Rahim, Muhammad, Yasir Akhtar, Miin-Shen Yang, Hago EM Ali, and Azhari A. Elhag. "Improved COPRAS method with unknown weights under p, q-QuasirungOrthopair Fuzzy environment: Application to green supplier selection." IEEE Access (2024).
37. Karnapke, Reinhardt, and Joerg Nolte. "Copra–a communication processing architecture for wireless sensor networks." In Euro-Par 2006 Parallel Processing: 12th International Euro-Par Conference, Dresden, Germany, and August 28–September 1, 2006. Proceedings 12, pp. 951-960. Springer Berlin Heidelberg, 2006.
38. Deb, ParthaPratim, Diptendu Bhattacharya, Indranath Chatterjee, Prasenjit Chatterjee, and EdmundasKazimierasZavadskas. "An intuitionistic fuzzy consensus WASPAS method for assessment of open-source software learning management systems." Informatica 34, no. 3 (2023): 529-556.
39. PK Kanumarlapudi. “Improving Data Market Implementation Using Gray Relational Analysis in Decentralized Environments” Journal of Artificial intelligence and Machine Learning., 2024, vol. 2, no. 1, pp. 1–7. doi: https://dx.doi.org/10.55124/ jaim.v2i1.271