Inquiry (E-Journal - Faculty of Business and Administration, International University of Sarajevo)
Not a member yet
209 research outputs found
Sort by
Implementation Of Long Short-Term Memory (LSTM) For User Authentication Based On Keystroke Dynamics
No full textThe paper reviews the usage, development and implementation of keystroke dynamics as a viable authentication system. AI methods are advancing, but we are still lacking biometric authentication systems in modern software which is being used daily. This paper shows the usage of long short-term memory layers for solving problems like keystroke dynamics and efficiently shows that with modern hardware, training and maintaining a small model is not taxing on the resources, as it may have been
Efficiency And Productivity Analysis In Turkish Banking Sector With Data Envelopment Analysis And Malmquist Index
Full text linkThe aim of this study is to analyze the efficiency and productivity of 3 public, 6 private and 6 foreign deposit banks operating in the Turkish banking sector with the help of data envelopment analysis and Malmquist index.For this purpose, the efficiency of 15 deposit banks operating in the Turkish banking sector between 2014 and 2018 was measured and whether the efficiency of Malmquist productivity index changed over the years. In the study, input-oriented Charnes Cooper and Rhodes (CCR) model was used under the assumption of constant return to scale and 4 input 2 output variables were selected. Inputs in efficiency and productivity measurement are defined as; personnel expenses / total assets (%), total loans / total assets (%), equity / total assets (%), total deposits / total assets (%), outputs are defined as; the earning power of assets (net profit / total assets), the earning power of equity (net profit / equity) (%). The Windows Data Envelopment Analysis Program (Win4Deap) package program was used in the analysis and brokerage approach was adopted. While 4 banks were active under the Constant Return to Scale(CRS) assumption between 2014 and 2018, 8 banks were found active under the assumption of Variable Return to Scale (VRS).Inefficient banks; target values were calculated by slacks movement and radial movement values to their original values and it was found that the lambda values calculated by Win4Deap program and which banks are peers. The changes observed in Malmquist total factor productivity, technical efficiency, technological efficiency, pure efficiency, scale efficiency and total factor productivity were analyzed as a whole and decision-making units experienced improvement in the 2016-2017 period
A De Novo Clustering Method: Snowball for Assigning 16S Operational Taxonomic Units
Full text linkTo analyze complex biodiversity in microbial communities, 16S rRNA marker gene sequences are often assigned to operational taxonomic units (OTUs). The abundance of methods that have been used to assign 16S rRNA marker gene sequences into OTUs brings discussions in which one is better. Suggestions on having clustering methods should be stable in which generated OTU assignments do not change as additional sequences are added to the dataset is contradicting some other researches contend that the methods should properly present the distances of sequences is more important. We add one more de novo clustering algorithm, Rolling Snowball to existing ones including the single linkage, complete linkage, average linkage, abundance-based greedy clustering, distance-based greedy clustering, and Swarm and the open and closed-reference methods. We use GreenGenes, RDP, and SILVA 16S rRNA gene databases to show the success of the method. The highest accuracy is obtained with SILVA library
Genomic Signal Processing Techniques for Taxonomy Prediction
Full text linkTo analyze complex biodiversity in microbial communities, 16S rRNA marker gene sequences are often assigned to operational taxonomic units (OTUs). The abundance of methods that have been used to assign 16S rRNA marker gene sequences into OTUs brings discussions in which one is better. Suggestions on having clustering methods should be stable in which generated OTU assignments do not change as additional sequences are added to the dataset is contradicting some other researches contend that the methods should properly present the distances of sequences is more important. We add one more de novo clustering algorithm, Rolling Snowball to existing ones including the single linkage, complete linkage, average linkage, abundance-based greedy clustering, distance-based greedy clustering, and Swarm and the open and closed-reference methods. We use GreenGenes, RDP, and SILVA 16S rRNA gene databases to show the success of the method. The highest accuracy is obtained with SILVA library
The Impact of the Institutional Environment on the Use of Licensed Technology
Full text linkThe robustness of the institutional environment is a requisite factor for the growth and development of a firm. This study is focused on the impact of factors of institutional environment on SMEs’ acquisition and use of licensed technology from abroad. The independent variables considered as the factors of institutional environment are: financial institutions, regulatory institutions, infrastructure, and security, while the dependent variable is the use of licensed technology from abroad. Data from the manufacturing and the service sectors of the economies of Africa and the Middle East are collected from the database of the World Bank Enterprise Survey. The survey employs random sampling to select firms in each country. The firms are stratified based on the number of employees and the geographical region. Questionnaires are administered to firms from 2006 to 2018 through cross-sectional data collection method. By focusing on the scope of research on the two regions and SMEs, the sampled observations are scaled down from 136,887 to 33,977 firms in 53 countries. Although not all the Pearson correlation coefficients of the independent variables with the dependent variable are high, there are satisfactory levels of significance with p-values below 5%. The independent variables in the regression model have a statistically significant impact on the use of licensed technology from abroad. The forecasting power of the regression model, the possible implications from the test results are shown. The limitations of the research and the possible areas for future research are discussed in the last section
Comparison of Different Machine Learning Algorithms for Breast Cancer Recurrence Classification
Full text linkIn this paper we compared some machine learning algorithms to predict recurrence of breast cancer and see which model used gives best accuracy for the prediction. In this study we used database donated by University Medical Centre, Institute of Oncology, Ljubljana, Slovenia. The preprocessed dataset includes 286 instances, 9 attributes and 1 class attribute. Firstly, we used attribute evaluation to see which attribute is more effective on class attribute. Secondly we have explored three different algorithms: C4.5, Random Forest and K Nearest Neighbor. Several data mining tools have been applied with these 3 algorithms to explore which model is better on accuracy. Finally we have found that C4.5 algorithm is the best for our dataset: breast cancer recurrence
Deep Transfer Learning for Food Recognition
Full text linkFood Recognition is an essential topic in the area of computer of its target applications is to avoid achieving a cashier at the dining place. In this paper, we investigate the application of Deep Transfer Learning for food recognition. We fine-tune three well learning models namely; AlexNet, GoogleNet, and Vgg16. The fine tuning procedure depends on removing the last three layers of each model and adds another five new layers. The training and validation of each model conducted through food a dataset collected from our university's canteen. The dataset contains 39 food types, 20 images for each type. The fine-tuned models show similar training and validation performance and achieved 100% accuracy over the small-scale dataset
TYPE-2 FUZZY TOPSIS MODEL FOR GREEN THIRD PARTY LOGISTICS PROVIDER PERFORMANCE EVALUATION
Full text linkGreen supplier selection, along with the environmental dimension in the supply chain, has attracted great interest both in the academic and institutional framework however, the supplier selection as well as the evaluation the performance of the current supplier affects the performance of the company and is important. In the real life, uncertainties in decision-making process are an integral part of this process. are things that exist in the nature of decision-making. Fuzzy set theory with the linguistic preferences was used to transform subjective decision-maker perceptions into a tangible net value. In this paper, it is proposed an interval type-2 fuzzy TOPSIS approach for green performance evaluation in GSCM. Then, It is applied in the performance evaluation of 3rd Party Logistics (3PL) providers to validate the presented model
Investigation Of 16S rRNA Gene And Gene Segments For The Determination Of Probiotics
No full textThe 16s-rRNA consists of hypervariable regions (V1 – V9) that demonstrate considerable sequence diversity among different bacteria. Species-specific sequences within a given hypervariable region constitute useful targets for diagnostic assays and other scientific investigations. Usually the size of the gene region is 1500 bp, which is large enough to be analyzed using bioinformatic tools and applied for detection. The need to advance the knowledge of the 16s-rRNA gene segments in bacterial strains would allow better understanding and better diagnostic possibilities when dealing with them. This could also be the basis for investigation of pathogenic microorganisms
Transfer Learning Utilization for Banknote Recognition: a Comparative Study Based on Bosnian Currency
No full textTransfer learning introduces the ability to perform deep learning models over a small set of data. This paper investigates the utilization of three fine-tuned Convolutional Neural Networks (CNNs), namely, Alexnet, Googlenet, and Vgg16. Alexnet and Googlenet consider as the state-of-the-art models in deep learning, while Vgg16 preference due to its depth. Each model was fine-tuned, trained, and tested over a dataset contains Bosnian Banknotes (BAM). The dataset covers 11 classes where 10 images were collected through mobile phone camera for each class. Alexnet showed a better performance in terms of completing the training while Vgg16 showed better performance in terms of accuracy as it achieved 100% compared to 95.24% for Alexnet. Googlenet showed less efficient performance by achieving 88.65%