YEDITEPE UNIVERSITY

Software Engineering Research Group

SERG

NEWS

How Is E-commerce Shaped by AI-Powered Recommendation Systems?

Nika Golestani, Eylül Akar, Ulaş Can Demirbağ, M. Akif Şahin, Buğrahan Aydın, Samet Alper Özdemir, Berk Tahir Kılıç


(3rd May 2024)

The importance of recommendation systems for e-commerce platforms is growing today. As Yeditepe Software Engineering Research Group, with our mission to develop sustainable solutions, we have focused on designing AI-powered recommendation systems. These systems have the potential to improve customer experience while also increasing sales.


How Does Our Project Work?

Our project offers personalized product recommendations by analyzing users' past data. Our work is modeled using various machine learning algorithms and big data analytics. The system of our project consists of the following modules:


1. AI-Powered Basket Recommendation System: This module analyzes users' order histories to recommend complementary products that can be added to the cart. The recommendations are optimized both for the user and complementary products, aiming to maximize sales potential.


2. Spell Check Module: Corrects spelling errors made in the product search bar on B2B platforms. This module detects entered incorrect words and offers correct word suggestions, thus we expect to increase the accuracy of search results.


3. Customer Categorization Module: This module segments customers based on criteria such as shopping frequency, expenditures, and the date of last purchase. Segmentation is anticipated to make marketing strategies more targeted and effective.


Lastly, our AI-powered recommendation system, successfully implemented on Eryaz Information Technologies' B2B e-commerce platform, aims to significantly improve customer experience and increase sales. This system has been developed using technologies such as pandas, mlxtend, Levenshtein distance, numpy, and matplotlib. Thanks to these technologies, we are able to perform complex operations on data and provide dynamic recommendations.


#EryazSoftware #ArtificialIntelligence #TensorFlow #Keras #MachineLearning #TechnologyInnovation

ERYAZ SOFTWARE
AI powered Recommendation System

Ulaş Can Demirbağ, M. Akif Şahin, Buğrahan Aydın, Nika Golestani, Eylül Akar, Samet Alper Özdemir, Berk Tahir Kılıç


(29th March 2024)

We are making significant progress in the artificial intelligence-supported recommendation system project that we developed specifically for Eryaz Software. This system is built using advanced technologies such as TensorFlow and Keras and is designed to provide the most appropriate recommendations to our customers' users.

The AI-powered recommendation system you are developing is designed to meticulously analyze user behavior through big data analytics and machine learning algorithms. This detailed analysis forms the basis for providing personalized recommendations. By doing this, the system aims to significantly improve the user experience and increase the efficiency of business operations. Your approach at Eryaz involves addressing this through several additional modules, with the work being modular and evolving to be adaptable to any type of data. This method not only provides flexibility in implementation, but also enables the system to evolve with changing needs and market dynamics, making it a robust and future-proof solution.

As a result, the AI-supported recommendation system can increase sales by providing personalized recommendations, encourage repeat purchases by strengthening customer loyalty, optimize inventory management with more accurate demand forecasts, improve marketing strategies by analyzing user behavior, provide competitive advantage and support strategic decision-making processes based on data. can make it happen. With these features, the system can significantly contribute to the profitability and long-term success of the company by improving the efficiency of business operations while improving the customer experience.

It is a source of great excitement for us to share the development process of our project and how it will create value for Eryaz Software's customers. Stay tuned to follow the development of our artificial intelligence-supported recommendation system and be with us on this exciting journey.


#EryazSoftware #ArtificialIntelligence #TensorFlow #Keras #MachineLearning #TechnologyInnovation

Dijital Dünya Çalıştayı (Digital World Workshop)

21 Aralık 2023

“Dijital Dünya Çalıştay” etkinliği DFDS’in organizatörlüğü ve sponsorluğunda 21 Eylül 2023 tarihinde DFDS Ataşehir Ar-ge merkezinde gerçekleştirilmiştir. Çalıştay tüm gün sürmüş olup gün boyu araştırmacı konuşmaları, tez sunumları, öğrenci sunumları ve tartışma grup etkinlikleri gerçekleştirilmiştir. Çalıştaya, DFDS Ar-ge merkezi mühendisleri, Yeditepe Üniversitesi Yazılım Mühendisliği Araştırma Grubu öğrencileri, ve bazı akademisyenler katılım göstermiştir.

Çalıştay kapsamında, 2 akademisyen uzman olduğu araştırma alanlarında konuşmalar vermiştir. Yeditepe Üniversitesi Bilgisayar Mühendisliği öğretim üyesi Doç.Dr. Mert Özkaya “C4 temelli Yazılım Mimarisi Tasarımı” üzerine bir konuşma vermiştir. Ayrıca, Yeditepe Üniversitesi Bilgisayar Mühendisliği öğretim üyesi Doç.Dr. Dionysis Goularas “Convolutional Neural Networks” üzerine bir konuşma verip katılımcıları bilgilendirmiştir.

Çalıştay kapsamında, Ece Gökçen Sarı isimli öğrenci “Dijital İkizler” üzerine bir sunum gerçekleştirmiştir. Bu sunum kapsamında, Ece dijital ikiz kavramını katılımcılara tanıtıp bu yönde var olan teknik ve teknolojileri, endüstrinin dijital ikizlere bakış açılarını ve dijital ikiz tasarlayıp geliştirmek için izlenmesi gereken yol haritasını aktarmıştır.

Nilhan Süer ve Ceren Yazgan isimli öğrenciler ise “Akıllı Tarım” üzerine bir sunum gerçekleştirdiler. Bu sunum kapsamında, öğrenciler akıllı tarımın tarihçesini, akıllı tarım sistemleri ve bu sistemlerin endüstrideki yerini, akıllı tarım sistemlerini tasarlayıp geliştirmek için önem verilmesi gereken araştırma alanlarını ve izlenmesi gereken yol haritasını aktarmışlardır.

Asude Demirboğa, Ece Tipici, ve Berke Özkeleş isimli öğrenciler de “Dijital Ekosistemler” üzerine bir sunum gerçekleştirdiler. Bu sunumda, öğrenciler dijital ekosistemin ne anlama geldiğini, endüstride var olan dijital ekosistem örneklerini, dijital ekosistemin faydalarını, endüstrinin dijital ekosistemlere bakış açısını ve şirketlerin kendi dijital ekosistemlerini oluşturmaları için izlenmesi gereken yolu aktardılar.

Ayrıca, 2 son sınıf öğrenci arkadaşımız, lisans bitirme tezlerini anlatan sunumlar gerçekleştirdiler.

Son olarak, 60 dakikalık paralel tartışma oturumları düzenlendi. Her bir oturumda DFDS mühendisleri, akademisyenler ve öğrenciler yer aldı ve yuvarlak bir masa etrafında oturup tahta eşliğinde problem-odaklı tartışmalara dahil oldular. Bu oturumlarda, DFDS’in yaşadığı çeşitli problemler mühendisler tarafından ortaya atıldı ve tüm oturum katılımcıları beraberce alternatif çözümleri tartıştılar. Katılımcıların hemfikir olduğu çözümlerin birçoğu projeleştirilip öğrencilere bitirme tezi olarak verildi. Ayrıca, bazı proje önerilerinin not edilip ileride TÜBİTAK destek programlarına uygun formatta dokümante edilmesi kararlaştırıldı.

YEDITEPE-SERG TALK 
Digital Ecosystems

Berke Özkeleş, Ece Tipici and Asude Beyaz Demirboğa

(20th December 2023)

Digital ecosystems are examined in this presentation by Asude Beyza Demirboğa, Berke Özkeleş, and Ece Tipici. It begins with a description of these ecosystems and their intricate relationships. The reasons why a digital ecosystem is important are discussed in the next section, with special attention on how it may help people and organizations work together more effectively and be more innovative and adaptable. The final segment explores how to create a robust digital ecosystem through practical measures like stakeholder involvement, technological integration, and strategic planning. The conclusion includes real-world examples of effective digital ecosystem deployments. In today's fast-paced technology context, this overview offers a clear understanding that enables individuals and organizations to grasp, support, and implement digital ecosystems.

YEDITEPE-SERG TALK 
Exploring Digital Agriculture

Ceren Yazgan and Nilhan Süer

(7th December 2023)

Nilhan Süer and Ceren Yazgan recently had an insightful presentation on Digital Farming at Yeditepe University. The presentation commenced with a comprehensive exploration of the Evolution of Agriculture from 1.0 to 4.0. They defined Smart Farming, explaining its significance in optimizing agricultural processes through technology integration. They explored the distinctive agricultural challenges in Turkey and talked about the digital solutions.

The presentation explored the indispensable role of Computer Science Engineering Professions in Digital Farming. Software Architecture was discussed, emphasizing its application in digital farming and the need for this technology in enhancing overall system efficiency. The presentation examined Digital Twin and Decision Support Systems, clarifying their importance in constructing virtual models to facilitate well-informed decision-making. The presentation also delved into the explanation of utilizing Domain Specific Languages for modeling, highlighting how they are applied to customize solutions to meet the specific demands of digital farming. Lastly, the integration of Artificial Intelligence and IoT in agriculture was explored, emphasizing their transformative impact on farm management.

In the conclusion, Nilhan and Ceren provided a global and local perspective on applications in Digital Farming. They highlighted specific applications developed in Turkey and globally, offering a well-rounded understanding of the technological advancements shaping the future of agriculture.

YEDITEPE-SERG TALK 
Explore the Power of Digital Twin 

Ece Gökçen Sarı

(16th November 2023)

In the delivered presentation, the speaker adeptly explored the realm of digital twins . Digital twins' historical roots were explained, taking participants back to the sources of inspiration and offering a basic understanding of their evolutionary process.Using material from the seminal article "Systems Architecture Design Pattern Catalog for Developing Digital Twins" by Bedir Tekinerdoğan and Cor Verdouw, the speaker skillfully navigated the topic of digital twins in the presentation that was given. The study of Tekinerdoğan and Verdouw was particularly noteworthy as a foundation for comprehending the origins and development of digital twins. The talk focused on how software design plays a crucial part in determining the characteristics of digital twins. Important passages from the cited article were woven into the story in a seamless manner, illustrating design patterns that act as tactical manuals for developers and architects creating digital twins. The discussion centered on the critical role that software architecture plays in defining the properties of digital twins. Essential sections from the referenced article were skillfully integrated into the narrative, showcasing design patterns that serve as practical guides for developers and architects building digital twins.

September 2023

During our discussion today, we delved into the intricacies of the article titled 'Data Fusion Analysis and Synthesis Framework for Improving Disaster Situation Awareness'.

Our main points of focus revolved around the concepts of data fusion and the utilization of future diagrams. We explored how these elements contribute to enhancing our understanding and preparedness in the face of disasters, providing valuable insights into the framework's potential applications in real-world scenarios.

September 2023

Meeting on Potential Collaboration Opportunities with Prof. Amir Kumar Mishra from University of Cape Town

In a recent meeting at Yeditepe University Department of Computer Engineering, we had the pleasure of hosting Professor Amir Kumar Mishra from the University of Cape Town, along with Professor Mert Ozkaya, the head of Yeditepe University Software Engineering Research Group. Our discussions centered around the exciting field of software engineering and its potential to tackle industry-wide challenges.

During our conversation, we emphasized the critical role of sensor technologies in software systems, highlighting their significance in acquiring data that is essential for making informed decisions. The quality of this data was stressed, underlining the need for accurate and noise-free information.

Our discussions underscored the importance of sensor technologies in shaping the future of software engineering, enabling precise and reliable data acquisition essential for the success of various applications across diverse industries.



The meeting concluded with a shared enthusiasm for future collaborations. Professor Mert Ozkaya expressed a keen interest in exploring projects that contribute to the broader field of software engineering.

Stay tuned for updates on Yeditepe University Software Engineering Research Group's website as we embark on an exciting journey to make strides in overcoming industry challenges through the innovative applications of software engineering. Together, we are poised to advance the field and create solutions that benefit society as a whole.

July 2023

A meeting about Decision Support Systems

We delved into the depths of Decision Support Systems (Karar Destek Sistemleri) architecture during our meeting. These systems play a crucial role in facilitating the most efficient and prompt actions. For instance, in the context of fire detection systems, the process was previously carried out manually.

One intriguing aspect explored during our meeting was the development process of these systems. Traditionally, the focus has been on coding, but the research highlighted the importance of considering architectural software development. The architecture of Decision Support Systems is a critical factor in their overall efficiency and effectiveness. Yet, at present, this aspect remains relatively underexplored. Participants expressed interest in further exploring the architectural aspects to ensure the seamless integration of these systems with existing infrastructure and processes, ultimately optimizing their performance and impact.

Let's also mention that we can showcase these systems in a simulator and generate source code from them. During our meeting, we discussed how the ability to demonstrate these systems in a simulator adds another layer of understanding and allows for practical experimentation without real-world consequences. Moreover, the capability to generate source code from these simulated models streamlines the implementation process, saving valuable time and effort for developers and ensuring a smoother transition from concept to reality. We were excited about the potential of this feature, as it opens up new possibilities for testing, refining, and deploying Decision Support Systems in a more agile and efficient manner.

The discussions during the meeting emphasized the need to move beyond merely coding the systems and instead, focus on defining their architectural design. By considering architectural software development, Decision Support Systems can be elevated to a new level of sophistication, leading to faster and more accurate decision-making processes.

July 2023

A Graphical Modeling Editor for Managing Objectives and Key Results (OKR)

We developed a graphical editor for the companies to manage their Objectives and Key Results (OKR).

With the editor, you may draw an organisational chart consisting of the business units and their dependencies and draw an OKR diagram for each business unit. The OKR diagram for a business unit is specified with the objectives and their key results and the associations of those objectives with the objectives of the dependent units.

Also, the editor can be used for checking a set of properties at modeling time automatically including the objectives with no key results and the objectives that are not associated with the objectives of the dependent units.

The editor can further produce some useful reports from the model specifications.

Note: The project motivation has been originated by DFDS Development Center in Istanbul and the tool has been developed using Eclipse Sirius by Cenk Cöne as a graduation project. 

July 2023

Dr. Ozkaya travelled to Rome, Italy to present the conference paper entitled "Practitioners’ Experiences on Developing Graphical Modeling Editors: A Survey" at the ICSOFT conference.

The paper can be accessible here

June 2023.

We developed an architecture modeling language and its prototype toolset for the specification of Decision Support Systems and their simulations.

Please see the project web-site

" We developed a data mapping modeling language (March 2023). "

" We conducted DFDS Workshop on 30 January 2023. " 

" Dr. Mert Özkaya gave a talk respect to Model Based Engineering. "