Informatica

The article focuses on the presentation and comparison of selected heuristic algorithms for solving the inverse problem for the anomalous diffusion model. Considered mathematical model consists of time-space fractional diffusion equation with initial boundary conditions. Those kind of models are used in modelling the phenomena of heat flow in porous materials. In the model, Caputo’s and Riemann-Liouville’s fractional derivatives were used. The inverse problem was based on identifying orders of the derivatives and recreating fractional boundary condition. Taking into consideration the fact that inverse problems of this kind are ill-conditioned, the problem should be considered as hard to solve. Therefore,to solve it, metaheuristic optimization algorithms popular in scientific literature were used and their performance were compared: Group Teaching Optimization Algorithm (GTOA), Equilibrium Optimizer (EO), Grey Wolf Optimizer (GWO), War Strategy Optimizer (WSO), Tuna Swarm Optimization (TSO), Ant Colony Optimization (ACO), Jellyfish Search (JS) and Artificial Bee Colony (ABC). This paper presents computational examples showing effectiveness of considered metaheuristic optimization algorithms in solving inverse problem for anomalous diffusion model.

Neural networks (NNs) are well established and widely used in time series forecasting due to their frequent dominance over other linear and nonlinear models. Thus, this paper does not question their appropriateness in forecasting cryptocurrency prices; rather, it compares the most commonly used NNs, i.e. feedforward neural networks (FFNNs), long short-term memory (LSTM) and convolutional neural networks (CNNs). This paper contributes to the existing literature by defining the appropriate NN structure comparable across different NN architectures, which yields the optimal NN model for Bitcoin return forecasting. Moreover, by incorporating turbulent events such as COVID and war, this paper emerges as a stress test for NNs. Finally, inputs are carefully selected, mostly covering macroeconomic and market variables, as well as different attractiveness measures, the importance of which in cryptocurrency forecasting is tested. The main results indicate that all NNs perform the best in an environment of bullish market, where CNNs stand out as the optimal models for continuous dataset, and LSTMs emerge as optimal in direction forecasting. In the downturn periods, CNNs stand out as the best models. Additionally, Tweets, as an attractiveness measure, enabled the models to attain superior performance.

Spam reviews are a pervasive problem on online platforms due to its significant impact on reputation. However, research into spam detection in data streams is scarce. Another concern lies in their need for transparency. Consequently, this paper addresses those problems by proposing an online solution for identifying and explaining spam reviews, incorporating data drift adaptation. It integrates (i) incremental profiling, (ii) data drift detection & adaptation, and (iii) identification of spam reviews employing Machine Learning. The explainable mechanism displays a visual and textual prediction explanation in a dashboard. The best results obtained reached up to 87% spam F-measure.

Artificial Intelligence (AI) in the price management process is being applied in business practice and research to a variety of pricing use cases that can be augmented or automated, providing opportunities as a forecasting tool or for price optimization. However, the complexity of evaluating the technology to prioritize implementation is challenging, especially for small and medium enterprises (SMEs), and guidance is sparse. Which are the relevant stakeholder criteria for a sustainable implementation of AI for pricing purpose? Which type of AI supported price functions meet these criteria best? Theoretically motivated by the hedonic price theory and advances in AI research, we identify nine criteria and eight AI supported price functions (AISPF). A multiple attribute decision model (MADM) using the fuzzy Best Worst Method (BWM) and fuzzy combined compromise solution (CoCoSo) is set up and evaluated by pricing experts from Germany and Spain. To validate our results and model stability, we carried out several random sensitivity analyses based on the weight of criteria exchange. The results suggest accuracy and reliability as the most prominent attribute to evaluate AISPF, while ethical and sustainable criteria are sorted as least important. The AISPF which best meet the criteria are financial prices followed by procurement prices.

Ontologies are used to semantically enrich different types of information systems (IS), ensure a reasoning on their content and integrate heterogeneous IS at the semantical level. On the other hand, fuzzy theory is employed in IS for handling the uncertainty and fuzziness of their attributes, resulting in a fully fuzzy IS. As such, ontology- and fuzzy-based IS (i.e. ontology and fuzzy IS) are being developed. So, in this paper, we present a bibliometric analysis of the ontology and fuzzy IS concept to grasp its main ideas, and to increase its body of knowledge by providing a concept map for ontology and fuzzy IS. The main results obtained show that by adding ontologies and fuzzy theory to traditional ISs, they evolve into intelligent ISs capable of managing fuzzy and semantically rich (ontological) information and ensuring knowledge recognition in various fields of application. This bibliometric analysis would enable practitioners and researchers gain a comprehensive understanding of the ontology and fuzzy IS concept that they can eventually adopt for development of intelligent IS in their work.

This work introduces ALMERIA, a decision-support tool for drug discovery. It estimates compound similarities and predicts activity, considering conformation variability. The methodology spans from data preparation to model selection and optimization. Implemented using scalable software, it handles large data volumes swiftly. Experiments were conducted on a distributed computer cluster using the DUD-E database. Models were evaluated on different data partitions to assess generalization ability with new compounds. The tool demonstrates excellent performance in molecular activity prediction (ROC AUC: 0.99, 0.96, 0.87), indicating good generalization properties of the chosen data representation and modelling. Molecular conformation sensitivity is also evaluated.

In recent years, Magnetic Resonance Imaging (MRI) has emerged as a prevalent medical imaging technique, offering comprehensive anatomical and functional information. However, the MRI data acquisition process presents several challenges, including time-consuming procedures, prone motion artifacts, and hardware constraints. To address these limitations, this study proposes a novel method that leverages the power of generative adversarial networks (GANs) to generate multi-domain MRI images from a single input MRI image. Within this framework, two primary generator architectures, namely ResUnet and StarGANs generators, were incorporated. Furthermore, the networks were trained on multiple datasets, thereby augmenting the available data, and enabling the generation of images with diverse contrasts obtained from different datasets, given an input image from another dataset. Experimental evaluations conducted on the IXI and BraTS2020 datasets substantiate the efficacy of the proposed method compared to an existing method, as assessed through metrics such as Structural Similarity Index (SSIM), Peak Signal-to-Noise Ratio (PSNR) and Normalized Mean Absolute Error (NMAE). The synthesized images resulting from this method hold substantial potential as invaluable resources for medical professionals engaged in research, education, and clinical applications. Future research gears towards expanding experiments to larger datasets and encompassing the proposed approach to 3D images, enhancing medical diagnostics within practical applications.

The development of various digital social network platforms has caused public opinion to play an increasingly important role in the policy making process. However, due to the fact that public opinion hotspots usually change rapidly (such as the phenomenon of public opinion inversion), both the behaviour feature and demand feature of netizens included in the public opinion often vary over time. Therefore, this paper focuses on the feature identification problem of public opinion simultaneously considering the multiple observation time intervals and key time points, in order to support the management of policy-focused online public opinion. According to the variable-scale data analysis theory, the temporal scale space model is established to describe candidate temporal observation scales, which are organized following the time points of relevant policy promulgation (policy time points). After proposing the multi-scale temporal data model, a temporal variable-scale clustering method (T-VSC) is put forward. Compared to the traditional numerical variable-scale clustering method, the proposed T-VSC enables to combine the subjective attention of decision-makers and objective timeliness of public opinion data together during the scale transformation process. The case study collects 48552 raw public opinion data on the double-reduction education policy from Sina Weibo platform during Jan 2023 to Nov 2023. Experimental results indicate that the proposed T-VSC method could divide netizens that participate in the dissemination of policy-focused public opinion into clusters with low behavioural granularity deviation on the satisfied observation time scales, and identify the differentiated demand feature of each netizen cluster at policy time points, which could be applied to build the timely and efficient digital public dialogue mechanism.

Multi-agent approach is very popular for modelling and simulation of complex phenomena, design and programming of decentralised computing systems. Asynchronous beings, which do not share state but communicate using messages are a convenient abstraction for representing various phenomena observed in the real world. When the number of the considered agents grows, the designers and developers of such systems must address the problem of performance. Introducing distribution is often a weapon of choice, which, however, does not guarantee obtaining proper scalability and efficiency. The intensity of communication in a large-scale agent-based system can easily exceed the abilities of a distributed hardware architecture, leading to poor performance. After analysing various distributed agent-based systems, we identified several reasons for limited performance and several architectural solutions, which can help overcoming this problem. The main aim of the presented work is identification and systematization of these architectural solutions in the form of design patterns. As a result, we propose three new design patterns for building scalable distributed agent-based systems. A systematic description of their aims, structure, variants and features is provided, together with examples of applications.

The Multi-Objective Mixed-Integer Programming (MOMIP) problem is one of the most challenging. To derive its Pareto optimal solutions one can use the well-known Chebyshev scalarization and Mixed-Integer Programming (MIP) solvers. However, for a large-scale instance of the MOMIP problem, its scalarization may not be solved to optimality, even by state-of-the-art optimization packages, within the time limit imposed on optimization. If a MIP solver cannot derive the optimal solution within the assumed time limit, it provides the optimality gap, which gauges the quality of the approximate solution. However, for the MOMIP case, no information is provided on the lower and upper bounds of the components of the Pareto optimal outcome. For the MOMIP problem with two and three objective functions, an algorithm is proposed to provide the so-called interval representation of the Pareto optimal outcome designated by the weighting vector when there is a time limit on solving the Chebyshev scalarization. Such interval representations can be used to navigate on the Pareto front. The results of several numerical experiments on selected large-scale instances of the multi-objective multidimensional 0–1 knapsack problem illustrate the proposed approach. The limitations and possible enhancements of the proposed method are also discussed.