Computer Science
Loading...
ISSN 1508-2806
e-ISSN: 2300-7036
Issue Date
2024
Volume
Vol. 25
Number
No. 2
Description
Journal Volume
Computer Science
Vol. 25 (2024)
Projects
Pages
Articles
Sentiment-aware enhancements of PageRank-based citation metric, Impact Factor, and h-index for ranking the authors of scholarly articles
(Wydawnictwa AGH, 2024) Gupta, Shikha; Kumar, Animesh
Heretofore, the only way to evaluate an author has been frequency-based citation metrics that assume citations to be of a neutral sentiment. However, considering the sentiment behind citations aids in a better understanding of the viewpoints of fellow researchers for the scholarly output of an author. We present sentiment-enhanced alternatives to three conventional metrics namely Impact Factor, h-index, and PageRank-based index. The proposal studies the impact of the proposed metrics on the ranking of authors. We experimented with two datasets, collectively comprising almost 20,000 citation sentences. The evaluation of the proposed metrics revealed a significant impact of sentiments on author ranking, evidenced by a weak Kendall coefficient for the Author Impact Factor and h-index. However, the PageRank-based metric showed a moderate to strong correlation, due to its prestige-based attributes. Furthermore, a remarkable Rank-biased deviation exceeding 28% was seen in all cases, indicating a stronger rank deviation in top-ordered ranks.
Explainable Spark-based PSO clustering for intrusion detection
(Wydawnictwa AGH, 2024) Ben Ncir, Chiheb Eddine; Ben Haj Kacem, Mohamed Aymen; Alattas, Mohammed
Given the exponential growth of available data in large networks, the existence of rapid, transparent, and explainable intrusion detection systems has become of highly necessity to effectively discover attacks in such huge networks. To deal with this challenge, we propose a novel explainable intrusion detection system based on Spark, Particle Swarm Optimization (PSO) clustering, and eXplainable Artificial Intelligence (XAI) techniques. Spark is used as a parallel processing model for the effective processing of large-scale data, PSO is integrated to improve the quality of the intrusion detection system by avoiding sensitive initialization and premature convergence of the clustering algorithm and finally, XAI techniques are used to enhance interpretability and explainability of intrusion recommendations by providing both micro and macro explanations of detected intrusions. Experiments are conducted on large collections of real datasets to show the effectiveness of the proposed intrusion detection system in terms of explainability, scalability, and accuracy. The proposed system has shown high transparency in assisting security experts and decision-makers to understand and interpret attack behavior.
Finding the inverse of a polynomial modulo in the ring Z[x] based on the method of undetermined coefficients
(Wydawnictwa AGH, 2024) Yakymenko, Ihor; Kasianchuk, Mykhailo; Karpinski, Mikolaj; Shevchuk, Ruslan; Shylinska, Inna
This paper presents the theoretical foundations of finding the inverse of a polynomial modulo in the ring Z[x] based on the method of undetermined coefficients. The use of the latter makes it possible to significantly reduce the time complexity of calculations avoiding the operation of finding the greatest common divisor. An example of calculating the inverse of a polynomial modulo in the ring Z[x] based on the proposed approach is given. Analytical expressions of the time complexities of the developed and classical methods depending on the degrees of polynomials are built. The graphic dependence of the complexity of performing the operation of finding the inverse of a polynomial in the ring Z[x] is presented, which shows the advantages of the method based on undetermined coefficients. It is found that the efficiency of the developed method increases logarithmically with an increase in the degrees of polynomials.
Detection of credit card fraud with optimized deep neural network in balanced data condition
(Wydawnictwa AGH, 2024) Shome, Nirupam; Sarkar, Devran Dey; Kashyap, Richik; Laskar, Rabul Hussain
Due to the huge number of financial transactions, it is almost impossible for humans to manually detect fraudulent transactions. In previous work, the datasets are not balanced and the models suffer from overfitting problems. In this paper, we tried to overcome the problems by tuning hyperparameters and balancing the dataset with a hybrid approach using under-sampling and oversampling techniques. In this study, we have observed that these modifications are effective in getting better performance in comparison to the existing models. The MCC score is considered an important parameter in binary classification since it ensures the correct prediction of the majority of positive data instances and negative data instances. So, we emphasize on MCC score and our method achieved an MCC score of 97.09%, which is far more (16 % approx.) than other state-of-the-art methods. In terms of other performance metrics, the result of our proposed model has also improved significantly.
Clustering for clarity: improving word sense disambiguation through multilevel analysis
(Wydawnictwa AGH, 2024) Dubey, ShivKishan; Kohli, Narendra
ExistingWord Sense Disambiguation (WSD) techniques have limits in exploring word-context relationships since they only deal with the effective use of word embedding, lexical-based information via WordNet, or the precision of clustering algorithms. In order to overcome this limitation, the study suggests a unique hybrid methodology that makes use of context embedding based on center-embedding in order to capture semantic subtleties and utilizing with a two-level k-means clustering algorithm. Such generated context embedding, producing centroids that serve as representative points for semantic information inside clusters. Additionally, go with such captured cluster- centres in the nested levels of clustering process, locate groups of linked context words and categorize them according to their word meanings that effectively manage polysemy/ homonymy as well as detect minute differences in meaning. Our proposed approach offers a substantial improvement over traditional WSD methods by harnessing the power of center-embedding in context representation, enhancing the precision of clustering and ultimately overcoming existing limitations in context-based sense disambiguation.