• International Journal of Computer Science & Network Security
• /
• v.21 no.9
• /
• pp.179-184
• /
• 2021

In our study we performed a sentiments analysis from the images. For this purpose, we used 153 images that contain: people, animals, buildings, landscapes, cakes and objects that we divided into two categories: images that suggesting a positive or a negative emotion. In order to classify the images using the two categories, we created two models. The SAI-G model was created with Google's AutoML Vision service. The SAI-C model was created on the Clarifai platform. The data were labeled in a preprocessing stage, and for the SAI-C model we created the concepts POSITIVE (POZITIV) AND NEGATIVE (NEGATIV). In order to evaluate the performances of the two models, we used a series of evaluation metrics such as: Precision, Recall, ROC (Receiver Operating Characteristic) curve, Precision-Recall curve, Confusion Matrix, Accuracy Score and Average precision. Precision and Recall for the SAI-G model is 0.875, at a confidence threshold of 0.5, while for the SAI-C model we obtained much lower scores, respectively Precision = 0.727 and Recall = 0.571 for the same confidence threshold. The results indicate a lower classification performance of the SAI-C model compared to the SAI-G model. The exception is the value of Precision for the POSITIVE concept, which is 1,000.

• ETRI Journal
• /
• v.43 no.2
• /
• pp.299-312
• /
• 2021

It is necessary to achieve high performance in the task of zero anaphora resolution (ZAR) for completely understanding the texts in Korean, Japanese, Chinese, and various other languages. Deep-learning-based models are being employed for building ZAR systems, owing to the success of deep learning in the recent years. However, the objective of building a high-quality ZAR system is far from being achieved even using these models. To enhance the current ZAR techniques, we fine-tuned a pretrained bidirectional encoder representations from transformers (BERT). Notably, BERT is a general language representation model that enables systems to utilize deep bidirectional contextual information in a natural language text. It extensively exploits the attention mechanism based upon the sequence-transduction model Transformer. In our model, classification is simultaneously performed for all the words in the input word sequence to decide whether each word can be an antecedent. We seek end-to-end learning by disallowing any use of hand-crafted or dependency-parsing features. Experimental results show that compared with other models, our approach can significantly improve the performance of ZAR.

• International Journal of Computer Science & Network Security
• /
• v.21 no.8
• /
• pp.79-86
• /
• 2021

Alzheimer's is a chronic progressive disease which exhibits varied symptoms and behavioural traits from person to person. The deterioration in cognitive abilities is more noticeable through their Activities and Instrumental Activities of Daily Living rather than biological markers. This information discussed in social media communities was collected and features were extracted by using the proposed fuzzy logic based algorithm to address the uncertainties and imprecision in the data reported. The data thus obtained is used to train machine learning models in order to predict the longevity of the patients. Models built on features extracted using the proposed algorithm performs better than models trained on full set of features. Important findings are discussed and Support Vector Regressor with RBF kernel is identified as the best performing model in predicting the longevity of Alzheimer's patients. The results would prove to be of high value for healthcare practitioners and palliative care providers to design interventions that can alleviate the trauma faced by patients and caregivers due to chronic diseases.

• International Journal of Computer Science & Network Security
• /
• v.22 no.7
• /
• pp.147-156
• /
• 2022

Ontologies are knowledge containers in which information about a specified domain can be shared and reused. An event happens within a specific time and place and in which some actors engage and show specific action features. The fact is that several ontology models are based on events called Event-Based Models, where the event is an individual entity or concept connected with other entities to describe the underlying ontology because the event can be composed of spatiotemporal extents. However, current event-based ontologies are inadequate to bridge the gap between spatiotemporal extents and participants to describe a specific domain event. This paper reviews, describes and compares the existing event-based ontologies. The paper compares various ways of representing the events and how they have been modelled, constructed, and integrated with the ontologies. The primary criterion for comparison is based on the events' ability to represent spatial and temporal extent and the participants in the event.

• International Journal of Computer Science & Network Security
• /
• v.22 no.6
• /
• pp.400-407
• /
• 2022

Innovations significantly affect the efficiency of the socioeconomic systems of the regions, acting as a system-forming element of their development. Modern models of economic development also consider innovation activity, intellectual potential, knowledge as the basic factors for stimulating the economic growth of the region. The purpose of the study is to develop methodological foundations for evaluating the effectiveness of a regional innovation system based on a multidimensional analysis of its effects. To further study the effectiveness of RIS, we have used one of the methods of multidimensional statistical analysis - canonical analysis. The next approach allows adding another important requirement to the methodological provision of evaluation of the level of innovation development of industries and regions, namely - the time factor, the formalization of which is realized in autoregressive dynamic economic and mathematical models and can be used in our research. Multidimensional Statistical Analysis for RIS effectiveness estimation was used to model RIS by typological regression. Based on it, multiple regression models were built in groups of regions with low and relatively high innovation potential. To solve the methodological problem of RIS research, we can also use the approach to the system as a "box" with inputs and outputs.

• The korean journal of orthodontics
• /
• v.52 no.2
• /
• pp.102-111
• /
• 2022

Objective: To develop and explore the usefulness of an artificial intelligence system for the prediction of the need for dental extractions during orthodontic treatments based on gender, model variables, and cephalometric records. Methods: The gender, model variables, and radiographic records of 214 patients were obtained from an anonymized data bank containing 314 cases treated by two experienced orthodontists. The data were processed using an automated machine learning software (Auto-WEKA) and used to predict the need for extractions. Results: By generating and comparing several prediction models, an accuracy of 93.9% was achieved for determining whether extraction is required or not based on the model and radiographic data. When only model variables were used, an accuracy of 87.4% was attained, whereas a 72.7% accuracy was achieved if only cephalometric information was used. Conclusions: The use of an automated machine learning system allows the generation of orthodontic extraction prediction models. The accuracy of the optimal extraction prediction models increases with the combination of model and cephalometric data for the analytical process.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.2
• /
• pp.438-455
• /
• 2024

Recently, many unsupervised learning-based models have emerged for Just Noticeable Difference (JND) estimation, demonstrating remarkable improvements in accuracy. However, these models suffer from a significant drawback is that their heavy reliance on handcrafted priors for guidance. This restricts the information for estimating JND simply extracted from regions that are highly related to handcrafted priors, while information from the rest of the regions is disregarded, thus limiting the accuracy of JND estimation. To address such issue, on the one hand, we extract the information for estimating JND in an Adversarial Complementary Learning (ACoL) way and propose an ACoL-JND network to estimate the JND by comprehensively considering the handcrafted priors-related regions and non-related regions. On the other hand, to make the handcrafted priors richer, we take two additional priors that are highly related to JND modeling into account, i.e., Patterned Masking (PM) and Contrast Masking (CM). Experimental results demonstrate that our proposed model outperforms the existing JND models and achieves state-of-the-art performance in both subjective viewing tests and objective metrics assessments.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.24 no.2
• /
• pp.224-232
• /
• 2000

Four turbulent $k-{\varepsilon}$ models (i.e., standard model, modified models with streamline curvature modification and/or preferential dissipation modification) are applied in order to analyze the turbulent flow of wall-attaching offset jet. For numerical convergence, this paper develops a method of slowly increasing the convective effect induced by skew-velocity in skew-upwind scheme (hereafter called Partial Skewupwind Scheme). Even though the method was simple, it was efficient in view of convergent speed, computer memory storage, programming, etc. The numerical results of all models show good prediction in first order calculations (i.e., reattachment length, mean velocity, pressure), while they show some deviations in ·second order (i.e., kinetic energy and its dissipation rate). Like the previous results obtained by upwind scheme, the streamline curvature modification results in better prediction, while the preferential dissipation modification does not.

• Journal of the Korean Mathematical Society
• /
• v.41 no.5
• /
• pp.821-864
• /
• 2004

Both parametric and parameter-free necessary and sufficient optimality conditions are established for a class of nondiffer-entiable nonconvex optimal control problems with generalized fractional objective functions, linear dynamics, and nonlinear inequality constraints on both the state and control variables. Based on these optimality results, ten Wolfe-type parametric and parameter-free duality models are formulated and weak, strong, and strict converse duality theorems are proved. These duality results contain, as special cases, similar results for minmax fractional optimal control problems involving square roots of positive semi definite quadratic forms, and for optimal control problems with fractional, discrete max, and conventional objective functions, which are particular cases of the main problem considered in this paper. The duality models presented here contain various extensions of a number of existing duality formulations for convex control problems, and subsume continuous-time generalizations of a great variety of similar dual problems investigated previously in the area of finite-dimensional nonlinear programming.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.25 no.1
• /
• pp.1-10
• /
• 2001

As computer power is increased, refined finite element models are employed for structural analysis. However, it is difficult and expensive to use refined models in the design stage. The refined models especially cause problems in the preliminary design where the design is frequently changed. Therefore, simplified models are needed. The simplification process is regarded as an empirical technique. Simplified and equivalent finite element model of a structure has been studied and used in the preliminary design. A general approach to establish the simplified and equivalent model is presented. The generated simple model has satisfactory correlation with the corresponding refined finite element model. An optimization method, the Goal Programming algorithm is used to make the simple model. The simplified model is used for the design change and the changed design is recovered onto the original design. The presented method was verified with three examples.