Vol. 1 No. 2 (2023): Vol. 1 No. 2 (2023)-- Publication of the second issue
We are excited to announce the upcoming release of the second issue of the esteemed journal, "Current Trends in Computing (CTC)." This signifies a significant milestone in our journey, further solidifying our commitment to providing a crucial platform for esteemed researchers and scholars. The forthcoming edition promises to build upon the success of our inaugural issue, presenting groundbreaking research and cutting-edge investigations within the dynamic field of computing.
In this second volume, readers can anticipate an impressive array of scholarly contributions, comprising not only a diverse selection of topics but also a testament to the continuous evolution of computing. We are proud to showcase the work of distinguished researchers who have delved into various facets of the discipline, unraveling new insights and presenting state-of-the-art advancements.
Within these pages, you will find a meticulously curated collection of articles, each contributing to the ever-expanding pool of knowledge in computing. This issue embodies our commitment to excellence and innovation, encapsulating the latest concepts, methodologies, and theories that define the forefront of computing research.
As we unveil the second issue of "Current Trends in Computing (CTC)," we extend an invitation to the research community and enthusiasts alike to join us on this transformative journey. Together, we will delve into the depths of cutting-edge research, exploring uncharted territories and unlocking limitless possibilities within the realm of computing.
Let this momentous occasion be a catalyst for collaboration, inspiration, and the collective pursuit of technological excellence. As we celebrate this milestone, we look forward to the shared discoveries and advancements that will shape the future of computing.Cover and Content
Please click here to access the articles in the second issue of the journal :
DEEP LEARNING-BASED HUMAN DETECTION FOR FALL INJURIES
Buse SARICAYIR, Esmanur ALICAN, Caner OZCAN
Abstract: Falls are a significant public health problem, especially among the elderly and people with limited mobility. A fall may seem like a minor accident, but the injuries that can result from a fall and the underlying health problems that can cause falls have a significant impact on people's lives. Especially in elderly individuals, such accidents occur more frequently and lead to more severe consequences. Research shows that one-third of homebound older adults and more than half of hospitalized older adults are at risk for falls. Falls can result in impaired balance and gait, fear of falling, disability, and a decline in daily activities and quality of life. This fear adversely affects the daily lives of elderly individuals. Therefore, real-time fall detection systems contribute to preventing more severe injuries. Our proposed method uses state-of-the-art deep learning techniques to detect and localize people in video streams. Its goal is to ensure the rapid provision of assistance to the person who has fallen after the incident. In the development stage of the paper, YOLOv7 and YOLOv8 architectures have been utilized. Furthermore, we discuss the potential applications of this approach in real-world scenarios, such as fall detection systems for elderly care, surveillance, and automated emergency response. The main contributions of this work are a novel deep learning approach to human detection in the context of fall injuries, practical applications of the proposed approach, and its potential to improve safety and quality of life for at-risk populations, especially the elderly and those with limited mobility.
FACE RECOGNITION APPROACH BY USING DLIB AND K-NN
Muhammed Taha AYDIN, Oğuzhan MENEMENCİOĞLU, İlhami Muharrem ORAK
Abstract: The face serves as a unique topographical map that reflects an individual's distinct features. Face recognition has gained prominence as a popular biometric method, especially in security control applications. In this study, we present a system developed using a Haar cascade classifier and Hog-based Dlib face detector for human face detection. Face features are extracted with the Dlib deep metric learning library, and classification is performed using the k-NN algorithm. The system underwent testing on benchmark data within the framework of an exam access control system. The system demonstrated an accuracy of up to 90% in the Orl_Face dataset. The measurement results were compared with other face recognition systems for validation. Beyond accuracy assessments, the proposed system was also compared with similar training tools, fostering a comprehensive discussion on its performance and capabilities.
TECHNIQUES USED TO EXTRACT FEATURES FROM CANDLESTICK CHARTS IN THE STOCK MARKET A SYSTEMATIC REVIEW
Edrees Ramadan MERSAL, Hakan KUTUCU
Abstract: In the financial sector, accurately forecasting stock market trends is essential for guiding the investment and trading decisions of investors and traders. These professionals often rely on candlestick charts to analyze and predict stock price fluctuations. In recent times, various methods and algorithms have been applied to leverage candlestick charts for prediction purposes. This systematic review aims to examine the application of Japanese candlesticks and machine learning techniques, including artificial neural networks, in predicting stock market trends. It also delves into the effective feature engineering strategies for extracting pertinent information from raw data, encompassing technical indicators and candlestick charts. The review encompasses 30 studies published between 2019 and 2023, selected based on criteria that include the utilization of candlestick charts in stock market analysis. The findings reveal that numerous studies employing automatic encoders, convolutional neural networks, and Gramian Angular Field (GAF) for feature geometry extraction from candlestick charts also identify common patterns.
TEXT CLASSIFICATION WITH FREQUENCY-BASED TEXT VECTORISATION METHODS FOR ENHANCING CALL CENTRE EFFICIENCY
Muammer ÖZDEMİR, Yasin ORTAKCI
Abstract: In today's business world, many transactions take place over the phone or online. Call centers play a significant role in dealing with different situations and solving problems that come with the large volume of global business. As an interface between companies/institutions and customers, call centers aim to eliminate problems, correct mistakes, resolve conflicts, and increase customer satisfaction. The traditional approach involves customer service agents handling inquiries and complaints, but human error can hinder effective problem resolution. Intelligent assistant applications have emerged to augment the skills of customer service agents, improve performance, and maximize customer satisfaction. This study focuses on addressing the challenges faced by the Republic of Turkiye Ministry of Trade Call Center in the (RTMTCC), which handles over 10,000 calls per day. For this purpose, it introduces an intelligent framework that uses AI-driven methods and frequency-based text vectorization techniques to efficiently route calls to relevant departments, with the aim of increasing customer satisfaction and reducing economic losses. Using historical call texts, Bag of Words (BoW), and Term Frequency-Inverse Document Frequency (TF-IDF), the study evaluates the performance of five different classifiers: Stochastic Gradient Descent (SGD), Logistic Regression (LR), Naive Bayes (NB), Adaptive Boosting (AdaBoost), Artificial Neural Networks (ANN). The results indicate that the AdaBoost classifier generally outperforms others in both text vectorization approaches by reaching higher precision, recall and f1-score values. The study provides new approaches to automate call routing, evaluates how to classify text effectively, and shows the strengths and weaknesses of different text analysis methods, helping us to understand call center operations better.
AUTOMATIC CLASSIFICATION OF WALNUT LEAF IMAGES WITH GRADCAM AND DEEP LEARNING
Alper Talha KARADENİZ , Erdal BAŞARAN, Yüksel ÇELİK
Abstract: Walnut leaves similar color and formation make distinguishing between varieties considerably challenging for individuals. Examining and categorizing such plant leaves one by one can be a time-consuming and costly process. Hence, experimental studies are conducted in laboratory settings to classify walnut varieties. Within the scope of this study, an original dataset consisting of 1751 walnut leaf images obtained from 18 different walnut varieties was prepared. Various preprocessing techniques were applied to the original dataset, and additionally, data augmentation methods were employed to obtain an expanded dataset. Both datasets were trained using deep learning models. Among these models, the Vgg16 CNN model demonstrated the most superior performance. The proposed model, trained with Vgg16 on the augmented dataset, produced Gradcam images and was further classified using the Vgg16 CNN algorithm. According to experimental test results, the proposed model achieved a success rate of 77.11%. This study demonstrates the successful utilization of deep learning techniques for classifying walnut varieties from walnut leaf images.
COMPARING THE PRACTICAL DIFFERENCES BETWEEN DECISION TREE AND RANDOM FOREST ALGORITHMS IN ESTIMATING HOUSING PRICES
Pınar SARI, İpek Doğa BEDİRHAN, Çağrı SEL
ABSTRACT: The process of estimating the price of houses is becoming increasingly important in light of the changing economy worldwide, as houses are considered a basic need and a source of investment. This process is aimed at preventing losses, market monitoring, minimizing problems, and arriving at accurate conclusions in the face of complex structures and issues. To achieve this, modern technology introduces the concepts of artificial intelligence and machine learning, which are integrated into all areas of life, to make progress in the process. Although machine learning and the algorithms used in this field have become widespread in recent years, there are still not enough studies on housing pricing. At the same time, people remain unaware of the field of machine learning and its applicability in every sector. Machine learning in general; It expands the data pool and enables new prediction results to be created by making future predictions based on data. Decision Tree Algorithm; In addition to facilitating understanding and interpretation in every field, it can handle multi-output problems and minimize preparation with its easy integration structure. Random Forest Algorithm can prevent the overfitting problem in classification problems and can be applied in both regression and classification problems. The aim of the study is to popularize the use of machine learning algorithms in the real estate sector. This will allow for effective housing price predictions during times of uncertainty and help in selecting the appropriate method by comparing the algorithms. Additionally, this study aims to reduce existing problems using these algorithms. A data set called "California Housing Prices," containing 20.640 samples and eight features, was used in this study. The results of the Decision Tree and Random Forest Algorithms were examined on this data set. Performance evaluation and comparison were made using MSE, RMSE, R2, and MAE metrics. It was observed that the Random Forest Algorithm produced better results and was superior to the Decision Tree Method when predicting house prices.