The initial literature reviewing step is of great importance during any scientific reporting. Nevertheless, finding relevant papers grows tough as the number of online scientific publications rapidly increases. Correspondingly, the need for article recommendation systems has emerged, which aim to recommend new papers suitable for the researchers’ interests. Using these systems provides researchers access to related publications quickly and effectively. In this study, a novel article recommendation system, which is empowered by the hybrid combinations of content-based state-of-the-art methods, is proposed. Various methods are utilized comparatively for an in-depth analysis, and user profiles are evaluated. 41,000 articles collected from the ARXIV dataset are used in the performance evaluation. In the experiments in which Word2vec and LDA are combined, Precision@50, Recall@50, and F1-score@50 achieve the highest performance with .206, .791, and .498 values, respectively. The in-depth analysis and the numerical findings justify that the proposed system is strong and promising compared to the literature.
The stage before the conversion of agricultural products into post-harvest consumer products is the process of separating the raw products into appropriate classes. Today, this difficult manual separating process is a process in which a large number of workers work at an intense pace on the product line and the workforce is intensively spent. Disruptions in separating as a result of carelessness cause product loss, loss of time and cost increases. In this study, as an alternative to manual separating processes, a real-time separating system, which detects the products in the factory band with object recognition methods and enables fast positioning of the separating tool on the products, works simultaneously with object recognition and traveling salesman problem algorithms has been created. In this way, a low-budget separating system is recommended for large selecting processes with a time- and cost-effective selecting model. In the study, the creation of a real-time fast separating system with the support of the traveling salesman algorithm, performance evaluation and research and findings on the fast separating model are presented.
Today, smart phones based on different mobile communication technologies such as GSM, GPRS/EDGE, 3G, 4G, 5G and 6G have become an indispensable element for various services such as communication, entertainment, and banking. In this context, determining the base station power level is important in terms of limit values and public health, especially in places with high power density. In this study, electromagnetic power density of 31 different base stations was measured at 900 MHz frequency at 20, 40 and 60 meters distances from base stations. Since it is practically not possible to measure each base station from every distance, 3 different distances were chosen randomly. Then, using the power density values measured from different distances, electromagnetic power density estimation was made with multiple linear regression (MLR) analysis method for distances at intermediate distances (25, 30, 35, 45, 50 and 55 meters). In this study, MLR analysis method was applied for power density estimation for the first time in the literature and adjusted R-square value above 0.99 was obtained for each intermediate distance.
In this research, a hybrid artificial neural network (ANN) optimized by a genetic algorithm (GA) was used to estimate energy and exergy analyses parameters. This article presents an approach for estimating energy and exergy analyses parameters with optimized ANN model based on GA (GA-ANN) for different ternary blends consisting of diesel, biodiesel and bioethanol in a single-cylinder, water-cooled diesel engine. The data used in the experiments performed at twelve different engine speeds between 1000 and 3000 rpm with 200 rpm intervals for five different fuel mixtures consisting of fuel mixtures prepared by blends biodiesel, diesel and 5% bioethanol in different volumes constitute the input data of the models. Using these input data, engine torque (ET), amount of fuel consumed depending on fuels and speed (AFC), carbon monoxide emission values (CO), carbon dioxide emission values (CO2), hydrocarbon emission values (HC), nitrogen oxides emission values (NOx), the amount of air consumed (AAC), exhaust gas temperatures (EGT) and engine coolant temperatures (ECT) were estimated with the GA-ANN. In examining the results obtained were examined, it was proved that diesel, biodiesel and bioethanol blends were effective in predicting all the results mentioned in engine studies performed at 200 rpm intervals in the 1000-3000 rpm range. A standard ANN model used in the literature was also proposed to measure the prediction performance of GA-ANN model. The predictive results of both models were compared using various performance indices. As a result, it was revealed that the proposed GA-ANN model reached higher accuracy in estimating the exergy and energy analyses parameters of the diesel engine compared to the standard ANN technique.
Many online service providers use a recommendation system to assist their customers' decision-making by generating recommendations. Accordingly, this paper proposes a new recommendation system for tourism customers to make online reservations for hotels with the features they need, saving customers time and increasing the impact of personalized hotel recommendations. This new system combined collaborative and content-based filtering approaches and created a new hybrid recommendation system. Two datasets containing customer information and hotel features were analyzed by Recency, Frequency, Monetary (RFM) method in order to identify customers according to their purchasing nature. The main idea of the recommendation system is to establish correlations between users and products and make the decision to choose the most suitable product or information for a particular user. As a result of the exponential growth of online data, this vast amount of information for use in the tourism industry can be leveraged by decision-makers to make purchasing decisions[20]. Filtering, prioritizing, and beneficially presenting relevant information reduces this overload. There are following three main ways that recommendation systems can generate a recommendation list for a user; content-based, collaborative-based, and hybrid approaches[1]. This paper describes each category and its techniques in detail. RFM Analysis is used to identify customer segments by measuring customers' purchasing habits. It is the process of labeling customers by determining the Recency, Frequency, and Monetary values of their purchases and ranking them on a scoring model. Scoring is based on how recently they bought (Recency), how often they bought (Frequency), and purchase size (Monetary). Experimental results show that the accuracy of behavior analysis using Manhattan distance-based hybrid filtering is greatly improved compared to collaborative and content-based algorithms.
As the world grows, the demand for transporting goods is increasing, the number of goods in factories and ports is increasing, to transport all these goods, cranes are indispensable. In fact, currently, crane rigs working in factories and ports operate with low stability, when working or the phenomenon of swaying of the load occurs, leading to inaccurate positioning, loss of safe transportation of goods. To overcome these shortcomings, the paper proposes the design of a neural-fuzzy adaptive controller combined with an LQR controller (ANFIS-LQR) to control the forklift's position in the shortest time to achieve the desired exact position. At the same time, we want to control the deflection angle of the load so that the vibration when working is minimal. To check and evaluate the quality and stability of the system; the proposed design controller is simulated on MATLAB/Simulink software in the case of changes in system parameters and noise affecting the gantry crane system. To evaluate the superiority of the paper compared with published works, the author compares ANFIS-LQR with other published control methods such as DE-PID, Fuzzy-PD, Fuzzy dual and Fuzzy sliding, the simulation results show that the neural-fuzzy adaptive controller combined with the proposed LQR controller works well t_xlvt=2.1s , t_xlgt=3.5s, 0max=0.3(rad).
The unmanned aerial vehicle name is TOMHA, which was developed to be used in the response of social incidents, aims to support the operational activities of security forces in response to social incidents, to expand their domination areas, to detect incidents that may disturb social peace in advance and to provide rapid intervention with the new unmanned aerial vehicle technologies developed. The scope of the TOMHA system designed was kept comprehensive compared to other unmanned aerial vehicles and the scope includes intervention to social events, ordering in local administrations, defense, reconnaissance and attack activities in military operations, inspections arranged for public interest, AFAD and service areas, forest fires detection and intervention, and public order operations. This TOMHA is being developed using the Pixhawk flight control card and the jetson nano artificial intelligence card. In addition to these cards, it has the feature of manual or artificial intelligence supported autonomous flight thanks to GPS, telemetry, FPV transceiver module, camera systems and national software to be used. It is controlled through the controller using RC communication channel for manual use. TOMHA has a flight time of 13.6 minutes, a thrust of 4.45G and a speed of 78 km and a mileage of 4993 meters in optimum conditions. The findings obtained by the tests performed with the designed TOMHA prototype show similar results with the literature. Thanks to the national design, TOMHA stands out when it encounters other unmanned aerial vehicles. It is seen that the response of the system to the sudden changes caused by the maneuver movements in the simulation environment is very fast and it follows the changes.
Due to the increase in the world population day by day, the amount of food needed is also increasing day by day. Diseases that occur in plants reduce the amount and quality of the product obtained. In this study, a computer-aided model was developed to detect diseases in tea leaves. Because plant diseases can be difficult and misleading to detect with the naked eye by farmers or experts. It is very important to detect diseases in tea leaves using artificial intelligence methods. Three Convolutional Neural Network (CNN) architectures accepted in the literature were used as the basis for the classification of diseases in tea leaves. With these three CNN architectures, feature maps of the images in the data set were obtained. After combining the feature maps obtained in each architecture, they were classified in the Linear Discriminant classifier. In addition, the performance of the proposed model was compared with seven CNN architectures accepted in the literature. The performance of the models used in the study was evaluated using different performance measurement metrics. The obtained results showed that the proposed model can be used to classify diseases in tea leaves.
In order to sustain human life without problems, a rational planning is required for the conservation and use of existing water resources. The potential of future water sources should be determined as the first step in such planning. Therefore, river flow forecasting is necessary to provide basic information about a variety of problems related to the operation of river systems. In this study, the long-term daily flow values of the Zamantı River-Değirmenocağı, Zamantı River-Ergenuşağı, and Eğlence River-Eğribük stations in the Seyhan Basin in Turkey were examined. In order to predict the forward flow rate from past flow measurement values, the Adaptative Neuro-Fuzzy Inference System (ANFIS) model was trained using Backpropagation (BP), Hybrid Learning (HB), and Simulated Annealing (SA) algorithms, and the results were compared. The performance of ANFIS models created with different input parameters using Grid Partitioning (GP) and Fuzzy C-Means Clustering (FCM) methods was also examined. The evaluation criteria used for comparison were Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Determination Coefficient (R2), and Mean Absolute Percentage Error (MAPE). The best results for R2 values of 0.6854, 0.9242, and 0.9373 were obtained for FMSs using the BP model. As a result of the analysis, it was concluded that the BP algorithm could be used more successfully and effectively than other algorithms for training ANFIS parameters in nonlinear problems.
Societies take various initiatives to reduce the impact of natural disasters. Unfortunately, certain nations and regions are better suited than others to finding solutions to the problem, whether for political, cultural, economic, or other factors. This paper deals with the cluster analysis of 170 countries based on world risk index and climate risk index data. We use the k-means approach for clustering in sequential stages of this work. Specifically, we first carry out both the elbow method and silhouette scores to determine the number of clusters. Then clustering analysis is carried out, taking into account the World Risk Index, which includes risks of both exposure and vulnerability. Second, the Climate Risk Index is implemented into the first stage results by clustering countries after determining the number of clusters. Lastly, statistical analyses on the change of clusters for exposure, vulnerability, and climate risk are investigated and discussed in detail. Taken together, each of the risk elements like earthquake, tsunami, socioeconomic development, health care capability, etc. differs by nation. Clusters of countries with similar risks are reported. When the climate risk index is included in the evaluation, the number of clusters increases. The Climate Risk Index has been determined as a variable that cannot be ignored when countries are clustered according to their risk profiles.
Both academic and commercial interest in agricultural robots has increased recently. This is due to the fact that agricultural robots address significant issues such as seasonal labor shortages during harvest and the rising concern over environmentally friendly practices. Because of these, several distinct agricultural robots have already been created for a variety of purposes, with varying degrees of success, including monitoring, spraying, harvesting, transport, etc. As a result, agriculture automation became unfeasible and unprofitable. The purpose of this study is to provide a new methodology for multitasking in performing agriculture operations by designing a 4 Degrees of freedom (4DoF) robotic arm with a new mechanism that has a different configuration with 2 grippers. We did kinematics and kinetics calculations using Denavit-Hardenberg (D-H) method with Lagrangian mechanics, and with the help of Christoffel Symbols of the First Kind, also We used Robot Operating System (ROS) to provide potential solutions using it. It is easy to be paired with other open-source technologies, such as android or IoT technologies. The robot arm can work synchronously with other hardware, sensors, cameras, and agricultural machines concerning farming operations. In conclusion, we believe that this new configuration will open a door for agricultural tasks to be easily automated and achieved using robotic technologies.
Battery management systems (BMS) are becoming essential for all types of electric vehicles using battery packs. Various factors, such as battery temperature and balance, directly affect the life, safety, and efficiency of batteries used in vehicles. For security and robustness, these factors should be monitored and adjusted instantly. Today, battery management systems are constantly being developed using different production methods and algorithms. In the studies, calculations are made by measuring parameters such as temperature, current, current balance, load status, and health status of the battery cells, and the control of the battery group is provided with these calculations. Instant and continuous measurement and processing of all these data and the creation of a control algorithm according to the calculation result are possible with the use of powerful processors. FPGA is a processor that can provide the speed and functionality required for BMS. In the battery management system, the FPGA is responsible for receiving and processing all signals from the battery cells and producing results. It instantly processes the data from temperature, current, and voltage sensors and applies the control stage required for balancing. In addition, the charge and discharge capacity of the battery is calculated by instantly measuring the state of charge (SOC). SOC is of great importance in the battery management system to ensure the safety of the battery pack. Therefore, the SOC needs to be estimated accurately and in real-time. Thanks to its parallel processing capability, the FPGA can simultaneously read data from the sensors and perform related calculations. In this study, a versatile system design with real-time, high computational speed for BMS was carried out on FPGA. The voltage and current of an experimental battery based on the embedded system were monitored in real time in a simulation environment. Experimental results show that the instantaneous SOC estimation is successful, and the system returns instant results to the incoming sensor data. The use of FPGA as a management unit will provide significant advantages in BMS with its high operating speed, real-time monitoring, low power consumption, and re-programmability.
Detection of pre-seizure signs in epileptic signals may help patients to survive the seizure with minimal damage. This study aims to detect epileptic seizure patterns using EEG datasets of five patients. The signals' maximum power spectral density (PSD) and high-frequency oscillations (HFOs) signals are investigated. The PSDs of all patients' signals are calculated, and the subbands of the maximum PSD are examined. It is observed that 95%, 85%, 85%, 75%, and 80% of the channels of the five patients are in the sum of delta and theta subbands of the maximum PSD, respectively. All patients' maximum power frequency subbands of F4 and T3 channels included only delta and theta subbands. Furthermore, frequency increase rates of pre-ictal and ictal signals are investigated, and increasing PSDs of ripples and fast ripples are then calculated. A much higher-frequency ripple follows the low-frequency ripple in 75%, 75%, 65%, 85%, and 75% of the channels of the first, second, third, fourth, and fifth patients, respectively. For the pre-ictal data, a much higher frequency ripple is not seen, followed by a low-frequency ripple in 90%, 85%, 75%, 90%, and 90% of all channels of five patients, respectively. In addition, in this study, the frequency of signals is observed to be 80 Hz and above in the Fp2, C4, P4, O2, and Pz channels, which are common to all patients. Consequently, examining PSD and HFO signals ensures the detection of the differences between the data sets and detects the epileptic seizure patterns in all five patients.
Iron metal is the most widely used metal type. This metal, which is used in countless sectors, is processed in different ways and turned into steel. Since steel has a brittle structure compared to iron, defects may occur in the plates during the rolling process. Detection of these defects at the production stage is of great importance in terms of commercial and safety. Machine learning methods can be used in such problems for fast and high accuracy detection. For this purpose, using a dataset obtained from stainless steel surface defects in this study, classification processes were carried out to detect defects with four different machine learning methods. Logistic Regression (LR), Decision Tree (DT), Support Vector Machine (SVM) and Random Forest (RF) algorithms were used for classification processes. The highest classification accuracy was obtained from the 79.44% RF model. Correlation analysis was performed in order to analyze the effects of the features in the dataset on the classification results. It is thought that the classification accuracy of the proposed models is satisfactory for this challenging problem, but needs to be upgraded.
Speech recognition has several application areas such as human machine interaction, classification of phone calls by gender, voice tagging, STT, etc. Predicting gender from audio signals is a problem that is easy for humans to solve, difficult to solve by a computer. In this study, a model based on MFCC and classification with machine learning is proposed for gender estimation from Turkish voice signals. Within the scope of the study, 58 different series and films were examined and a new original dataset was created with 894 audio recordings consisting of 5 sec sections taken from them. Mel-frequency cepstral coefficients (MFCC) and spectrogram, which are frequently used in the literature, were used for feature extraction from audio data. The results were first evaluated separately using two features in one way. A hybrid feature vector was then created using two feature vectors. Different machine learning algorithms (LR, DT, RF, XGB etc.) were tested in the classification process and it was seen that the best accuracy was achieved in the hybrid model and logistic regression with 89%. Recall, precision and f-score values were obtained as 86.8%, 92% and 89.3%, respectively. The obtained test results revealed that the proposed model, together with the hybrid feature vector used, the original dataset and the classifier based on machine learning, showed classification success in terms of accuracy and was a stable and robust model.
The aim of this study is to develop a multi-sensor and low-cost biosensor system using Arduino Uno microcontroller board. With this system, heart rate and body temperature can be measured digitally. In this way, the vital values of those with health problems will be monitored regularly, and sudden changes in vital values will be detected immediately. While developing the system, Arduino Uno control card, heart rate sensor and temperature sensor were used. In addition, a software sensitive to detecting changes in heart rate and body temperature has been developed. The system created is open to future additions and improvements. The data obtained through the sensors on the system can be easily tracked by transferring them to the computer over wi-fi. At the same time, it can be improved by placing more sensors in the system according to the purpose of use. As a result of the tests, it has been seen that the system successfully transmits the information received from the sensors to the users in the desired time and gives an audible warning at the determined critical values. Thanks to this system, vital changes in heart rate and body temperature can be accessed remotely and instantly, and it is thought that this will contribute to the field of medicine.
Load balancing and task scheduling are the main challenges in Cloud Computing. Existing load balancing algorithms have a drawback in considering the capacity of virtual machines while distributing loads among them. The proposed algorithm works toward solving existing issues, such as fair load distribution, avoiding underloading and overloading, and improving response time. It implements best practices of Throttled load balancing algorithm and Equally Shared Current Execution algorithm. Virtual machines are selected based on the ratio of their bandwidth and load allocation count. Requests are sent to a Virtual Machine with higher bandwidth and lower load allocation count. Proposed algorithm checks for the availability of VM based on their capacity. This process is performed by selecting two VMs and comparing their vmWeight capacity. The one with the least vmWeight is selected. CloudAnalyst is used for simulation, response time evaluation, and resource utilization evaluation. The simulation result of the proposed algorithm is compared with three well-known load-balancing algorithms. These are Round Robin, Throttled Load balancing algorithm, and Enhanced Active Monitoring. Load-balancing Proposed Algorithm selects VMs based on their Algorithm. The proposed algorithm has improved over other algorithms in load distribution, response time, and resource utilization. All virtual machines in the data centers are loaded with a relatively equal number of tasks according to their capacity. This resulted in fair resource sharing and load distribution.
One of the main problems of the developing and changing world is air pollution. In addition to human causes such as population growth, increase in the number of vehicles producing exhaust emissions in line with the population, development of industry, natural causes such as forest fires, volcano eruptions and dust storms also play a role in increasing air pollution. Air pollution has become a bigger problem that reduces the quality of life of living beings and causes various lung and heart diseases due to reasons such as the growing proximity of settlements to industrial zones due to population growth, the increase in the number of individual vehicles, and zoning works carried out by ignoring air quality. Both international organizations and local authorities take various measures to control and prevent air pollution. In Turkey, necessary legal arrangements have been made within the scope of these measures and air quality monitoring stations have been established. The task of these stations is to measure pollutants such as PM10, CO, SO2 together with meteorological data such as air temperature, humidity, wind speed and direction. In this study, a prediction model for the future concentrations of PM10, CO and SO2 pollutants using the measurement data from three different air quality monitoring stations in Konya between January 2020 and January 2021 was realized with a multi-layer Long Short Term Memory (LSTM) artificial neural network. The Root Mean Square Deviation (RMSE) and Mean Absolute Percentage Error (MAPE) methods was used to calculate the performance of the study. As a result of the study, it is observed that the multi-layer LSTM architecture is more successful than the single-layer architecture.
In this paper, a system for reading meteorological data like temperature, humidity, air pressure, wind speed, wind direction, and rainfall at regular intervals, deployed in Selçuk University Alaeddin Keykubat Campus. The system also provides real-time images and video time-lapses of the campus sky. These data are made available to university people via a website and mobile applications for both iOS and Android. The website and mobile applications provide a clear experience for the users, also explaining the icons and terms used on the website. Users can access the system archive in graphical ways.
In this study, the inductor current ripple and output voltage ripple of a two-phase DC-DC boost converter, circuit performance is investigated according to the direct/inverse coupling effect of the coupled inductors. The coupled inductors are modeled with both power electronics circuit and electromagnetic modeling and by using finite element analysis software (FEA). However, the high frequency inductor designs generally use air-gapped composite ceramic ferrite cores and are designed with powder core (Kool Mu) core structures that eliminate air gap requirements. Thus, the fringe flux, which occurs in the air gaps in ferrite cores and reduces the useful flux, and the bad effects such as overheating and electromagnetic noises (EMI/EMC) in the air gaps in high frequency switching are also reduced. Especially in interleaved power converter designs, the performances of coupled inductors affect the output parameters of the power electronics circuit. Considering energy efficiency and more compact circuit topologies, the modeling and simulation approach of high-frequency inductors using finite element analysis software, which is emphasized in this study, popularly and scientifically guides power electronics circuit designers.
