Alongside the modern software development life cycle approaches, software testing has gained more importance and has become an area researched actively within the software engineering discipline. In this study, machine learning and deep learning-related software fault predictions were made through a data set named SFP XP-TDD, which was created using three different developed software projects. A data set of five different classifiers widely used in the literature and their Rotation Forest classifier ensemble versions were trained and tested using this data set. Numerous publications in the literature discussed software fault predictions through ML algorithms addressing solutions to different problems. Some of these articles indicated the usage of feature selection algorithms to improve classification performance, while others reported operating ensemble machine learning algorithms for software fault predictions. Besides, a detailed literature review revealed that there were few studies involving software fault prediction with DL algorithms due to the small sample sizes in the data sets and the low success rates in the tests performed on these datasets. As a result, the major contribution of this research was to statistically demonstrate that DL algorithms outperformed ML algorithms in data sets with large sample values via employing three separate software fault prediction datasets. The experimental outcomes of a model that includes a layer of recurrent neural networks (RNNs) were enclosed within this study. Alongside the aforementioned and generated data sets, the study also utilized the Eclipse and Apache Active MQ data sets in to test the effectiveness of the proposed deep learning method.

Identification and location of defects in software projects is an important task to improve software quality and to reduce software test effort estimation cost. In software fault prediction domain, it is known that 20% of the modules will in general contain about 80% of the faults. In order to minimize cost and effort, it is considerably important to identify those most error prone modules precisely and correct them in time. Machine Learning (ML) algorithms are frequently used to locate error prone modules automatically. Furthermore, the performance of the algorithms is closely related to determine the most valuable software metrics. The aim of this research is to develop a Majority Vote based Feature Selection algorithm (MVFS) to identify the most valuable software metrics. The core idea of the method is to identify the most influential software metrics with the collaboration of various feature rankers. To test the efficiency of the proposed method, we used CM1, JM1, KC1, PC1, Eclipse Equinox, Eclipse JDT datasets and J48, NB, K-NN (IBk) ML algorithms. The experiments show that the proposed method is able to find out the most significant software metrics that enhances defect prediction performance.

The climate crisis is one of the most significant challenges of the twenty-first century. The primary cause of high carbon emissions is industrial production that relies on carbon-based energy sources such as fuel oil, paraffin, coal, and natural gas. One of the effective methods to minimize carbon emissions originating from the use of energy resources is using recycling systems. A blockchain-based recycling platform was developed in this regard, adhering to the basic principles of Industry 4.0, which Robert Bosch GmbH and Henning Kagermann’s working group described as an industrial strategy plan at the Hannover Fair in 2013. Concurrently, the recycling platform has set up an infrastructure that combines blockchain, AI, and IoT technologies for recycling objects. An IoT-based smart device was developed to collect recyclable objects. Thanks to the embedded artificial intelligence software and QR code sensor on the device, recyclable objects can be collected in different hoppers. In the laboratory studies, correct object recognition success was achieved at a rate of 98.2%.

Günümüzde insan kulak görüntülerinin kullanımı, biyometrik yetkilendirme ve gözetleme sistemlerinin sürdürülebilirliği adına önem kazanmaktadır. Güncel çalışmalar, böyle işlemlerin el ile yapılması yerine yarı otomatik veya tam otomatik olarak yapılabileceğini göstermektedir. Derin öğrenme soyut öznitelikleri (temsili öğrenme) kullanması nedeniyle klasik yöntemlere göre oldukça yüksek başarım değerlerine ulaşmaktadır. Çalışmamızda insanların cinsiyetlerine göre tam otomatik olarak sınıflandırılmasında insan kulağı görüntülerinin kullanımına dayanan melez derin öğrenme tabanlı sinerjik bir cinsiyet tanıma yaklaşımı oluşturulmuştur. Melezleme yoluyla hem evrişimli sinir ağı bileşeni hem de tekrarlayan sinir ağı tipli bileşenlerini bir arada içeren melez derin sinir ağı mimari modelleri kullanılmıştır. Bu modellerde tekrarlayan sinir ağı tipi bileşenler olarak uzun kısa süreli bellek ve kapılı tekrarlayan birim alınmıştır. Bu bileşenler sayesinde melez model görüntü içerisindeki piksel bölgeleri arasındaki ilişkisel bağımlılıkları oldukça iyi elde etmektedir. Bu sinerjik yaklaşım sayesinde çalışmamızdaki tek başına evrişimli sinir ağı modeline göre melez modellerin cinsiyet sınıflandırma doğruluğu daha yüksek olmaktadır. Cinsiyet işaretlemesine sahip iki farklı görüntü veri kümesi deneylerimizde kullanılmıştır. Deneysel sonuçların güvenirliği nesnel ölçütlerle kanıtlanmıştır. Yapılan deneylerde melez modellerle yapılan cinsiyet tanımada en yüksek değerler sırasıyla, EarVN veri kümesi için test doğruluğu %85,16 ve WPUT veri kümesi için test doğruluğu %87,61 oranlarında elde edilmiştir. Çalışmamızın son bölümünde tartışma ve sonuçlara yer verilmektedir. Nowadays, the use of the human ear images gains importance for the sustainability of biometric authorization and surveillance systems. Contemporary studies show that such processes can be done semi-automatically or fully automatically, instead of being done manually. Due to the fact that deep learning uses abstract features (i.e., representation learning), it reaches quite high performance values compared to classical methods. In our study, a synergistic gender recognition approach based on hybrid deep learning was created based on the use of human ear images in classifying people fully automatically according to their gender. By means of hybridization, hybrid deep neural network architectural models are used, which include both convolutional neural network component and recurrent neural network type components together. In these models, long-short term memory and gated recurrent unit are taken as recurrent neural network type components. Thanks to these components, the hybrid model extracts the relational dependencies between the pixel regions in the image very well. On account of this synergistic approach, the gender classification accuracy of hybrid models is higher than the standalone convolutional neural network model in our study. Two different image datasets with gender marking were used in our experiments. The reliability of the experimental results has been proven by objective metrics. In the conducted experiments, the highest values in gender recognition with hybrid models were obtained with the test accuracy of 85.16% for the EarVN dataset and 87.61% for the WPUT dataset, respectively. Discussion and conclusions are included in the last section of our study.

The text classification method is one of the most frequently used approaches in text mining studies. Text classification requires a model generation using a predefined dataset, and this model aims to assign uncategorized data to a correct category. In line with this purpose, this study used machine learning algorithms, deep learning algorithms, word embedding algorithms, and transfer-learning algorithms to classify Turkish texts using three diverse datasets, one of which is new, to analyze text classification performances for the Turkish language. The preparation process of the newly added dataset involved the variations in Turkish word usage patterns over the years, since it consisted of timestamp-enabled data. The study also developed a novel method named LSRM to increase the text classification performance for agglutinative languages such as Turkish. After testing the new method on datasets, the statistical ANOVA method revealed that applying the proposed LSRM method increased the classification performance.

The scattering number of a graph is a measure of the vulnerability of a graph. In this paper we investigate a refinement that involves the average of a local version of the parameter. If v is a vertex in a connected graph G, then s c v (G) = max { ω(G − S v ) − | S v | }, where the maximum is taken over all disconnecting sets S v of G that contain v. The average scattering number of G denoted by s c av (G), is defined as scav(G) = Σv ∈ V(G) scv(G) / n, where n will denote the number of vertices in graph G. Like the scattering number itself, this is a measure of the vulnerability of a graph, but it is more sensitive. Next, the relations between average scattering number and other parameters are determined. The average scattering number of some graph classes are obtained. Moreover, some results about the average scattering number of graphs obtained by graph operations are given.