Computer Engineering

Kalman filtreler (KF) özyineli kestiriciler olup, pek çok uygulama alanÕna sahiplerdir. Bu makalede KF kullanÕmÕnÕn tehdit (emiter) izleme amacÕyla bir grup ayrÕk olaylÕ iúlev üzerine kullanÕlmasÕndan bahsedilmektedir. Sistem ve zamanlama gürültülü birden fazla tehdide ait karmaúÕk (ham) darbe dizisi üzerinden tek bir tehdide ait olan darbe dizisini izlenmesi amaçlanmaktadÕr. KF'nin çalÕúabilmesi için gerekli parametreler, izleme algoritmasÕndan önce çalÕúan kimliklendirme algoritmasÕndan alÕnmaktadÕr. Geliútirilen KF, gürültüye ve kayÕp darbeler ile izlenen tehditten gelen darbeler arasÕndaki büyük zaman boúluklarÕna karúÕ da uyumludur. Algoritma, tehdide ait kimliklendirme aúamasÕnda tespit edilen frekans (TF), darbe geniúli÷i (DG), geliú açÕsÕ (GA), tarama örüntüsü (TO) gibi ayÕrÕcÕ parametreleri de içerdi÷i için sa÷lam ve verimli bir yapÕ sa÷lamaktadÕr.

COVID-19 is the most common infectious disease of the last few years and has caused an outbreak all around the world. The mortality rate, which was earlier in the hundreds, increased to thousands and then to millions. Since January 2020, several scientists attempted to understand and predict the spread of COVID-19 so that governments may make sufficient arrangements in hospitals and reduce the mortality rate. This research article presents a comparative performance analysis of ARIMA, Prophet and Holt-Winters Exponential Smoothing forecasting methods to make predictions for the COVID-19 disease epidemiology in Europe. The dataset has been collected from the World Health Organization (WHO) and includes the COVID-19 case data of European countries, which is categorized by WHO between the years of 2020 and 2022. The results indicate that Holt-Winters Exponential Smoothing method (RMSE: 0.2080, MAE: 0.1747) outperforms ARIMA and Prophet forecasting methods.

In recent years, the use of machine learning and data mining technologies has drawn researchers’ attention to new ways to improve the performance of Intrusion Detection Systems (IDS). These techniques have proven to be an effective method in distinguishing malicious network packets. One of the most challenging problems that researchers are faced with is the transformation of data into a form that can be handled effectively by Machine Learning Algorithms (MLA). In this paper, we present an IDS model based on the decision tree C4.5 algorithm with transforming simulated UNSW-NB15 dataset as a pre-processing operation. Our model uses Term Frequency.Inverse Document Frequency (TF.IDF) to convert data types to an acceptable and efficient form for machine learning to achieve high detection performance. The model has been tested with randomly selected 250000 records of the UNSW-NB15 dataset. Selected records have been grouped into various segment sizes, like 50, 500, 1000, and 5000 items. E...

❖ Employment of TF.IDF for transformation of network packet data for Machine Learning algorithms. ❖ An efficient model for distinguishing malicious network packets through the C4.5 algorithm. ❖ Remarkable improvements in the accuracy and performance with the combined use of TF.IDF and C4.5

A new hybrid machine learning method for the prediction of type 2 diabetes is introduced and explained in detail. Also outcomes are compared with the similar researches. Early prediction of diabetes is crucial to take necessary measures (i.e. changing eating habits, patient weight control etc.), to defer the emergence of diabetes and to reduce the death rate to some extent and ease medical care professionals' decision making in preventing and managing diabetes mellitus.The purpose of this study is the creation of a new hybrid feature selection approach combination of Correlation Matrix with Heatmap and Sequential forward selection (SFS) to reveal the most effective features in the detection of diabetes. A diabetes data set with 520 instances and seven features were studied with the application of the proposed hybrid feature selection approach. The evaluation of the selected optimal features was measured by applying Support Vector Machines(SVM), Random Forest(RF), and Artificial Neural Networks(ANN) classifiers. Five evaluation metrics, namely, Accuracy, F-measure, Precision, Recall, and AUC showed the best performance with ANN (99.1%), F-measure (99.1%), Precision (99.3%), Recall (99.1%), and AUC (99.8%). Our proposed hybrid feature selection model provided a more promising performance with ANN compared to other machine learning algorithms.

COVID-19 is the most common infectious disease of the last few years and has caused an outbreak all around the world. The mortality rate, which was earlier in the hundreds, increased to thousands and then to millions. Since January 2020, several scientists have attempted to understand and predict the spread of COVID-19 so that governments may make sufficient arrangements in hospitals and reduce the number of deaths. This research article presents a comparative performance analysis of ARIMA, Prophet and HoltWinters Exponential Smoothing forecasting methods to make predictions for the COVID-19 disease epidemiology in Europe. The dataset has been collected from the World Health Organization (WHO) and includes the COVID-19 case data of European countries, which is categorized by WHO between the years of 2020 and 2022. The results indicate that Holt-Winters Exponential Smoothing method (RMSE: 0.2080, MAE: 0.1747) outperforms ARIMA and Prophet forecasting methods.

Drinking water is becoming a crucial problem all over the world because of global warming. In crowded metropolises such as Istanbul, the problem of drinking water is a serious problem. In this study, it is aimed at developing a forecasting model by using the occupancy rates of Istanbul's dams between 2011 and 2020. The occupancy rate of each dam is then estimated using the best model for the years 2021 and 2022. According to the results of the estimation model, a decrease in the occupancy level of the dams in Istanbul is predicted. Therefore, it is thought that necessary measures should be taken to avoid water shortages.

In pursuit of better energy efficiency and enhanced network lifetime in Wireless Sensor Networks (WSNs), two crucial factors are data packet size and the transmission power level. Smaller packet size reduces the overall impact of bit error rates on packet loss, however, the consequence of smaller packet size is fragmentation into more data packets and thereby dissipation of increased energy. Hence, there emerges a delicate engineering tradeoff in deciding the data packet size where both low and high data packet size decisions lead to certain energy inefficiency issues. On the other hand, increasing transmission power level decreases packet loss probability, which is another decision variable to optimize for maximizing network lifetime. Joint consideration of these two factors exacerbates the complexity of the optimization problem for the objective of the network lifetime maximization. In this study, we utilized a realistic WSN link layer model built on top of empirically verified energy dissipation characteristics of Mica2 motes and WSN channel models. We make use of the aforementioned link layer model to form a novel Mixed Integer Programming (MIP) framework for joint optimization of transmission power level and data packet size to take up the challenge introduced above. Numerical evaluations of the MIP framework with discussion of the results over a large parameter space are performed to characterize the effects of joint optimization of packet size and power level on WSN lifetime.

❖ A novel Deep Learning-based model is proposed for detection of different types of fabric defects. ❖ Numerous models have tested and compared in the study in terms of both time and accuracy. ResNet50 and KNN outperform other models in fabric defect detection. ❖ Fabric defect detection is a critical issue in the textile industry. Deep Learning based systems are expected to reduce defect detection times with high accuracy potential.

Wireless sensor networks (WSNs) are widely used in various fields, and their deployment is critical to ensure area coverage and full network connectivity to achieve the maximum network lifetime. In this study, we present a mixed-integer programming (MIP) model that deeply investigates deployment parameters to optimize lifetime and analyze network connectivity. We further analyze the obtained results using Deep Belief Network (DBN) and Deep Neural Network (DNN) algorithms to achieve higher accuracy rates. Our evaluation shows that the DBN outperforms the DNN with an accuracy rate of 81.2%, precision of 81.2%, recall of 99.1%, and an F1-Score of 0.78. We also utilize two different datasets to justify the efficiency of the DBN in this research. The findings of this study emphasize the validity of our DBN algorithm and encourage further research into lifetime optimization and connectivity analysis in WSNs.

The text classification task has a wide range of application domains for distinct purposes, such as the classification of articles, social media posts, and sentiments. As a natural language processing application, machine learning and deep learning techniques are intensively utilized in solving such challenges. One common approach is employing the discriminative word features comprising Bag-of-Words and n-grams to conduct text classification experiments. The other powerful approach is exploiting neural network-based (specifically deep learning models) through either sentence, word, or character levels. In this study, we proposed a novel approach to classify documents with contextually enriched word embeddings powered by the neighbor words accessible through the trigram word series. In the experiments, a well-known web of science dataset is exploited to demonstrate the novelty of the models. Consequently, we built various models constructed with and without the proposed approach to monitor the models' performances. The experimental models showed that the proposed neighborhood-based word embedding enrichment has decent potential to be used in further studies.

k-means is the most widely used clustering algorithm due to its fairly straightforward implementations in various problems. Meanwhile, when the number of clusters increase, the number of iterations also tend to slightly increase. However there are still opportunities for improvement as some studies in the literature indicate. In this study, improved implementations of k-means algorithm with a centroid calculation heuristics which results in a performance improvement over traditional k-means are proposed. Two different versions of the algorithm for various data sizes are configured, one for small and the other one for big data implementations. Both the serial and MapReduce parallel implementations of the proposed algorithm are tested and analyzed using 2 different data sets with various number of clusters. The results show that big data implementation model outperforms the other compared methods after a certain threshold level and small data implementation performs better with increasing k value.

Choosing good candidates for the initial centroid selection process for compact clustering algorithms, such as k-means, is essential for clustering quality and performance. In this study, a novel hybrid evolutionary model for k-means clustering (HEkmeans) is proposed. This model uses meta-heuristic methods to identify the "good candidates" for initial centroid selection in kmeans clustering method. The results indicate that the clustering quality is improved by approximately 30% compared to the standard random selection of initial centroids. We also experimentally compare our method with the other heuristics proposed for initial centroid selection and the experimental results show that our method performs better in most cases.

The use of optimal ATM menu structuring for different customer profiles is essential because of usability, efficiency, and customer satisfaction. Especially in competitive industries such as banking, having optimal user interface (UI) is a must. Determining the optimal menu structure is generally accomplished through manual adjustment of the menu elements. However, such an approach is inherently flawed due to the overwhelming size of the optimization variables' search space. Previous studies on menu optimization either are based on customer questionnaires or made for only a specific menu type using heuristic approaches (i.e., not generic). In this paper, we propose an systematic optimization method for menu structuring problem through a novel Mixed Integer Programming (MIP) framework. Our optimization approach is not specific to a predetermined menu class, on the contrary, the MIP model is designed to be a generic optimization framework that can be applied to a wide range of menu optimization problems. We evaluated the performance gains on a dataset of actual ATM usage logs for a period of 18 months consisting of 40 million transactions. We validated our results with both simulation application and mining of existing data logs. The results show that the proposed optimization approach provides significant reduction in the average transaction completion time and the overall click count.

The use of optimal ATM menu structuring for different customer profiles is essential because of usability, efficiency, and customer satisfaction. Especially in competitive industries such as banking, having optimal user interface (UI) is a must. Determining the optimal menu structure is generally accomplished through manual adjustment of the menu elements. However, such an approach is inherently flawed due to the overwhelming size of the optimization variables' search space. Previous studies on menu optimization either are based on customer questionnaires or made for only a specific menu type using heuristic approaches (i.e., not generic). In this paper, we propose an systematic optimization method for menu structuring problem through a novel Mixed Integer Programming (MIP) framework. Our optimization approach is not specific to a predetermined menu class, on the contrary, the MIP model is designed to be a generic optimization framework that can be applied to a wide range of menu optimization problems. We evaluated the performance gains on a dataset of actual ATM usage logs for a period of 18 months consisting of 40 million transactions. We validated our results with both simulation application and mining of existing data logs. The results show that the proposed optimization approach provides significant reduction in the average transaction completion time and the overall click count.

k-means is the most widely used clustering algorithm due to its fairly straightforward implementations in various problems. Meanwhile, when the number of clusters increase, the number of iterations also tend to slightly increase. However there are still opportunities for improvement as some studies in the literature indicate. In this study, improved implementations of k-means algorithm with a centroid calculation heuristics which results in a performance improvement over traditional k-means are proposed. Two different versions of the algorithm for various data sizes are configured, one for small and the other one for big data implementations. Both the serial and MapReduce parallel implementations of the proposed algorithm are tested and analyzed using 2 different data sets with various number of clusters. The results show that big data implementation model outperforms the other compared methods after a certain threshold level and small data implementation performs better with increasing k value.