56271 research outputs found
Sort by
Production planning in industrial townships modeled as hub location-allocation problems considering congestion in manufacturing plants
No full textIn this paper, we aim to develop optimal production plans in industrial townships modeled as hub location-allocation problems (HLAP) taking congestion into account. In the proposed model, hub nodes are considered as industrial townships where manufacturing plants and a central distribution warehouse are located, and two objectives are targeted. The first is to minimize the total costs, which includes the cost of hub deployment, factories and warehouses, transportation, and so forth. The second is to minimize the total elapsed time of products in manufacturing plants and warehouses modeled as queues. Due to the ambiguity in estimating the model's parameters, they are considered as fuzzy parameters to make model closer to reality. The fuzzy model is then converted into an equivalent crisp model by combining the expected value (EV) and the fuzzy chance constrained programming (FCCP) approaches. Subsequently, the bi-objective crisp model is converted into a single aggregated objective model. In order to validate the proposed model, six numerical examples are solved, and the sensitivity of the proposed model with regard to changes in model's parameters is investigated
Development of a Bayesian Belief Network-based DSS for predicting and understanding freshmen student attrition
No full textStudent attrition – the departure from an institution of higher learning prior to the achievement of a degree or earning due educational credentials – is an administratively important, scientifically interesting and yet practically challenging problem for decision makers and researchers. This study aims to find the prominent variables and their conditional dependencies/interrelations that affect student attrition in college settings. Specifically, using a large and feature-rich dataset, proposed methodology successfully captures the probabilistic interactions between attrition (the dependent variable) and related factors (the independent variables) to reveal the underlying, potentially complex/non-linear relationships. The proposed methodology successfully predicts the individual students' attrition risk through a Bayesian Belief Network-driven probabilistic model. The findings suggest that the proposed probabilistic graphical/network method is capable of predicting student attrition with 84% in AUC – Area Under the Receiver Operating Characteristics Curve. Using a 2-by-2 investigational design framework, this body of research also compares the impact and contribution of data balancing and feature selection to the resultant prediction models. The results show that (1) the imbalanced dataset produces similar predictive results in detecting the at-risk students, and (2) the feature selection, which is the process of identifying and eliminating unnecessary/unimportant predictors, results in simpler, more understandable, interpretable, and actionable results without compromising on the accuracy of the prediction task
Efficacy and acceptability of psychosocial interventions in asylum seekers and refugees:systematic review and meta-analysis
No full textAimsIn the past few years, there has been an unprecedented increase in the number of forcibly displaced migrants worldwide, of which a substantial proportion is refugees and asylum seekers. Refugees and asylum seekers may experience high levels of psychological distress, and show high rates of mental health conditions. It is therefore timely and particularly relevant to assess whether current evidence supports the provision of psychosocial interventions for this population. We conducted a systematic review and meta-analysis of randomised controlled trials (RCTs) assessing the efficacy and acceptability of psychosocial interventions compared with control conditions (treatment as usual/no treatment, waiting list, psychological placebo) aimed at reducing mental health problems in distressed refugees and asylum seekers.MethodsWe used Cochrane procedures for conducting a systematic review and meta-analysis of RCTs. We searched for published and unpublished RCTs assessing the efficacy and acceptability of psychosocial interventions in adults and children asylum seekers and refugees with psychological distress. Post-traumatic stress disorder (PTSD), depressive and anxiety symptoms at post-intervention were the primary outcomes. Secondary outcomes include: PTSD, depressive and anxiety symptoms at follow-up, functioning, quality of life and dropouts due to any reason.ResultsWe included 26 studies with 1959 participants. Meta-analysis of RCTs revealed that psychosocial interventions have a clinically significant beneficial effect on PTSD (standardised mean difference [SMD] = ?irc;'0.71; 95% confidence interval [CI] ?irc;'1.01 to ?irc;'0.41; I 2 = 83%; 95% CI 78-88; 20 studies, 1370 participants; moderate quality evidence), depression (SMD = ?irc;'1.02; 95% CI ?irc;'1.52 to ?irc;'0.51; I 2 = 89%; 95% CI 82-93; 12 studies, 844 participants; moderate quality evidence) and anxiety outcomes (SMD = ?irc;'1.05; 95% CI ?irc;'1.55 to ?irc;'0.56; I 2 = 87%; 95% CI 79-92; 11 studies, 815 participants; moderate quality evidence). This beneficial effect was maintained at 1 month or longer follow-up, which is extremely important for populations exposed to ongoing post-migration stressors. For the other secondary outcomes, we identified a non-significant trend in favour of psychosocial interventions. Most evidence supported interventions based on cognitive behavioural therapies with a trauma-focused component. Limitations of this review include the limited number of studies collected, with a relatively low total number of participants, and the limited available data for positive outcomes like functioning and quality of life.ConclusionsConsidering the epidemiological relevance of psychological distress and mental health conditions in refugees and asylum seekers, and in view of the existing data on the effectiveness of psychosocial interventions, these interventions should be routinely made available as part of the health care of distressed refugees and asylum seekers. Evidence-based guidelines and implementation packages should be developed accordingly. Copyright © Cambridge University Press 2019 This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited
Businesspersons, governments and international law
Full text linkInternational law is a matter primarily for governments. Governments form and administer international law,1 and in this context, governments fascinatingly react to private commercial and financial initiatives. Businesspersons often challenge governments, and defining and conceptualizing this challenge is necessary to understand international law. United States President Donald Trump’s approach to international trade law should be seen in this light. He believes that the previous U.S. administrations did not make the necessary effort vis-à-vis corporations and foreign governments to favor the U.S. in foreign trade. President Trump believes that he is now giving the correct signals and incentives to businesspersons to contribute to the U.S. foreign trade and economy. A similar sort of tension exists in the approach of governments worldwide to the issue of cryptocurrency. Cryptocurrencies, e.g., Bitcoin, constitute a challenge to the current international financial and trade system as regulated by governments. Governments wish to control the development of cryptocurrencies through taxation, regulation, or even through prohibition. This Article sets forth two examples of evidence to substantiate the above argument: it starts by discussing international commercial law as a concept, with a focus on international commercial arbitration (“ICA”). This Article highlights the similarities between ICA and public international law dispute settlement. The Article then turns to cryptocurrencies, particularly bitcoin, the most important cryptocurrency for market capitalization
FETÖ'deki çözülmenin boyutları: Sebepler ve sonuçlar
No full textMuhalif mecraların tek bir çözüm teklifi üzerinde birleştikleri
söylenemez. Son gelinen noktada talep sahiplerini, Gülen’in yakın
çevresinin tasfiyesini isteyenler ile Gülen’in emekliliğini isteyenler
şeklinde iki gruba ayırabiliriz. FETÖ’nün “Hizmet” adı verilen görünür
örgütsel faaliyetine her iki grubun da karakteri itibarıyla hâlâ hayranlık
duyduğu ve bunun mutlaka canlandırılmasının gerektiği yönünde
müşterek bir kanaatin mevcut olduğu görülebilir
A fully integrated 2.4 dB NF capacitive cross coupling CG-LNA for LTE band
No full textThis paper presents a common gate low noise amplifier utilizing a passive feedback network that provides a competitive
and highly integrated front-end solution for mobile handset devices. This design utilizes a resistive load instead of the
inductive one used in other designs to reduce the on-chip silicon area. The design does not need an external matching
network which decrease the area of the PCB while achieving a sufficient input impedance matching, S11. It achieves a
measured gain higher than 20 dB, noise figure less than 3 dB and input referred third order intercept point (IIP3) value
higher than - 2.5 dBm at 2.3 GHz. The design is implemented in 65 nm UMC CMOS technology, occupies a total area of
0.065 mm2 and consumes 5 mW from a 1.4 V supply
Debt financing and the failure of innovation companies: The application of the CHS Model in U.S. stock markets
No full textThis study aims to determine whether increasing a firm's leverage significantly changes its level of bankruptcy risk in the innovative industry by using the CHS model [Campbell, J. Y., J. Hilscher, and J. Szilagyi. 2008. "In Search of Distress Risk." The Journal of Finance 63 (6): 2899-2939] to test 395 American innovation companies. These companies are categorised into four groups based on their debt ratios and their performance on the NYSE and NASDAQ stock exchanges is analysed in three separate periods. The findings reveal that innovation companies with a higher debt level are no riskier than those with a lower debt level
A deep learning approach to sentiment analysis in Turkish
Full text linkSentiment analysis is an application of natural language processing (NLP) which is a subfield of artificial intelligence. Sentiment analysis is used to determine the polarity of the thoughts mostly on social media posts, product or different media reviews. Due to its growing demand by data scientists and social media analysts it is one of the most popular topics in NLP. Beside the lexicon-based techniques, from well-known machine learning techniques to advanced algorithms such as deep learning algorithms, there are different kind of algorithms and approaches developed to obtain a good sentiment analysis tool.
This study proposes using recurrent neural networks, a type of deep learning algorithm for sentiment analysis in Turkish. Traditional machine learning methods such as logistic regression or Naive Bayes are often applied to this problem however their applicability is limited since they use bag-of-words model which does not take into account the order of the words in a sentence.
In this study we compare these approaches with a modern technique called recurrent neural networks using LSTM units on a dataset crawled from a Turkish movie website. Our results show that RNN based approaches improve the classification accuracies.Declaration of Authorship ii
Abstract iv
Öz v
Acknowledgments vii
List of Figures x
List of Tables xii
Abbreviations xiii
1 Introduction 1
2 Background 4 2.1 Early Work - Lexicon-based Approaches . . . . . . . . . . . . . . . . . . . 5 2.2 Machine Learning Approaches . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.2.1 Naive Bayes Method . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.2.2 Logistic Regression Method . . . . . . . . . . . . . . . . . . . . . . 6 2.2.3 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.3 Deep Learning Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.3.1 Recurrent Neural Networks (RNNs) . . . . . . . . . . . . . . . . . 9 2.3.2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.4 Practical Aspects of Running a Sentiment Analyzer . . . . . . . . . . . . . 10 2.4.1 Preprocessing - Text Vectorizers . . . . . . . . . . . . . . . . . . . 10 2.4.1.1 Tf-idf Vectorization of Sentences . . . . . . . . . . . . . . 10 2.4.1.2 Word Vector Representation . . . . . . . . . . . . . . . . 10 2.4.2 Hyperparameter Search . . . . . . . . . . . . . . . . . . . . . . . . 11 2.4.3 Measuring the Accuracy of Sentiment Analysis . . . . . . . . . . . 12
3 Dataset 13 3.1 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3.2 Data Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 3.3 Preprocessing Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3.3.1 Text Normalization . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3.3.2 Stemmer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3.3.3 Vectorization of Sentences . . . . . . . . . . . . . . . . . . . . . . . 17
3.3.3.1 Tf-idf Vectorizer . . . . . . . . . . . . . . . . . . . . . . . 17 3.3.3.2 Word Vector Representation . . . . . . . . . . . . . . . . 18
4 Experiments 19 4.1 Baseline Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 4.1.1 Naive Bayes Classifier . . . . . . . . . . . . . . . . . . . . . . . . . 20 4.1.2 Logistic Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 4.2 Deep Learning Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 4.3 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 4.4 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5 Conclusion 34 5.1 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 5.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Bibliography 3
Metamorfik zararlı yazılımların derin öğrenme ile sınırlandırılması
No full textIncreasing and widespread internet services and the users of these services are exposed to many malicious attacks. Although these attacks have many different purposes, they are generally done using malicious software. For this reason, too many malware are produced with different architecture and features. In the examinations conducted, it is seen that the competencies and activities of these malware have reached a very serious level. In this respect, metamorphic malware is the most advanced member of malware family. Metamorphic malware cannot be detected by anti-virus applications using traditional signature-based detection methods. As a result of this situation, it is not possible to classify these software according to their types. In this respect, almost all of the recent detection and classification studies address the behavior of malware. In this study, it is aimed to develop a classification method according to malware types by considering malware behavior. First of all, in our study, a dataset was created containing API calls made on the Windows operating system, which represents the behavior of malicious software. The data set contains the behavior of real malware such as Adware, Backdoor, Downloader, Dropper, Spyware, Trojan, Virus and Worm. Long-term term memory (LSTM), which is a widely used and deep learning method, is used as the classification method. In this way, a classification method has been developed by modeling the behaviors of 7 different types of malware.Yazarlk Beyan ii
Öz iv
Te³ekkür v
ekil Listesi viii
Tablo Listesi x
Ksaltmalar xii
1 Giri³ 1
2 lgili Çal³malar 5
3 Ön Bilgiler 10 3.1 Windows API Ça§rlar . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 3.2 Cuckoo Sandbox (Kum Havuzu) Uygulams . . . . . . . . . . . . . . . . . 11 3.3 Virus Total Servisi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 3.4 Derin Ö§renme - Deep Learning . . . . . . . . . . . . . . . . . . . . . . . . 14 3.4.1 Aktivasyon Foksiyonu . . . . . . . . . . . . . . . . . . . . . . . . . 14 3.4.2 LSTM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4 Yöntem 17 4.1 Veri Kümesi Olu³turma . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 4.2 Model Olu³turma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5 Deneyim 22 5.1 Snandrma Modellerinin lklendirilmesi . . . . . . . . . . . . . . . . . . 22 5.2 kili Snandrma Analiz Sonuçlar . . . . . . . . . . . . . . . . . . . . . . 23 5.2.1 kili LSTM Analiz Sonuçlar . . . . . . . . . . . . . . . . . . . . . . 24 5.2.1.1 Adware Zararl Yazlm Snf için Sonuçlar . . . . . . . . 26 5.2.1.2 Backdoor Zararl Yazlm Snf için Sonuçlar . . . . . . . 29 5.2.1.3 Downloader Zararl Yazlm Snf için Sonuçlar . . . . . . 32 5.2.1.4 Dropper Zararl Yazlm Snf için Sonuçlar . . . . . . . . 35 5.2.1.5 Spyware Zararl Yazlm Snf için Sonuçlar . . . . . . . . 38 5.2.1.6 Trojan Zararl Yazlm Snf için Sonuçlar . . . . . . . . . 41 5.2.1.7 Virus Zararl Yazlm Snf için Sonuçlar . . . . . . . . . 44
5.2.1.8 Worm Zararl Yazlm Snf için Sonuçlar . . . . . . . . . 47 5.2.1.9 Özet Sonuçlar . . . . . . . . . . . . . . . . . . . . . . . . 50 5.2.2 kili Geleneksel Yöntemler Analiz Sonuçlar . . . . . . . . . . . . . 50 5.2.2.1 Adware Zararl Yazlm Snf için Sonuçlar . . . . . . . . 50 5.2.2.2 Backdore Zararl Yazlm Snf için Sonuçlar . . . . . . . 51 5.2.2.3 Downloader Zararl Yazlm Snf için Sonuçlar . . . . . . 52 5.2.2.4 Dropper Zararl Yazlm Snf için Sonuçlar . . . . . . . . 53 5.2.2.5 Spyware Zararl Yazlm Snf için Sonuçlar . . . . . . . . 54 5.2.2.6 Trojan Zararl Yazlm Snf için Sonuçlar . . . . . . . . . 55 5.2.2.7 Virus Zararl Yazlm Snf için Sonuçlar . . . . . . . . . 56 5.2.2.8 Worm Zararl Yazlm Snf için Sonuçlar . . . . . . . . . 57 5.2.3 Derin Ö§renme ve Geleneksel Yöntemler Analiz Sonuçlar . . . . . 58 5.3 Çoklu Snandrma Analiz Sonuçlar . . . . . . . . . . . . . . . . . . . . . 59 5.3.1 Tek Katmanl LSTM Analiz Sonuçlar . . . . . . . . . . . . . . . . 59 5.3.2 ki Katmanl LSTM Analiz Sonuçlar . . . . . . . . . . . . . . . . . 62 5.3.3 DT Analiz Sonuçlar . . . . . . . . . . . . . . . . . . . . . . . . . . 66 5.3.4 kNN Analiz Sonuçlar . . . . . . . . . . . . . . . . . . . . . . . . . 67 5.3.5 RF Analiz Sonuçlar . . . . . . . . . . . . . . . . . . . . . . . . . . 68 5.3.6 SVM Analiz Sonuçlar . . . . . . . . . . . . . . . . . . . . . . . . . 69 5.3.7 Derin Ö§renme ve Geleneksel Yöntemler Çoklu Analiz Sonuçlar . 71
6 Sonuç 73
Kaynaklar 7