Expert Systems with Applications最新文献

{"title":"A multi-input lightweight convolutional neural network for breast cancer detection considering infrared thermography","authors":"Yundong Tang ,&nbsp;Depei Zhou ,&nbsp;Rodolfo C.C. Flesch ,&nbsp;Tao Jin","doi":"10.1016/j.eswa.2024.125738","DOIUrl":"10.1016/j.eswa.2024.125738","url":null,"abstract":"<div><div>Although deep convolutional neural network (CNN) has been widely used in the breast cancer detection based on thermal imaging technology, this scenario still did not receive enough attention in the mobile devices with limited resource. In addition, there still exists challenge on how to assist front view thermal imaging by side one during breast cancer detection. This study proposes a multi-input lightweight CNN named Multi-light Net in order to achieve more accurate early detection for breast cancer, which combines the thermal image from multiple perspectives with the lightweight CNN on the basis of model performance and scale. In addition, a new weighted label smoothing regularization (WLSR) is proposed for the Multi-light Net with the purpose of increasing the network’s generalization ability and classification accuracy. The experimental results demonstrate that the proposed approach by combining front view with side view can achieve more significant results than the common one using only front view during breast cancer detection, and the proposed Multi-light Net also exhibits an excellent performance with respect to the currently popular lightweight CNN. Furthermore, the proposed WLSR loss function can also lead to both faster convergence rate and more stable training process during network training and ultimately higher diagnostic accuracy for breast cancer.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125738"},"PeriodicalIF":7.5,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662274","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MIAN: Multi-head Incongruity Aware Attention Network with transfer learning for sarcasm detection","authors":"Xin Guan ,&nbsp;Jiuxin Cao ,&nbsp;Hui Zhang ,&nbsp;Biwei Cao ,&nbsp;Bo Liu","doi":"10.1016/j.eswa.2024.125702","DOIUrl":"10.1016/j.eswa.2024.125702","url":null,"abstract":"<div><div>Sarcasm is a common rhetorical metaphor in social media platforms, that individuals express emotion contrary to the literal meaning. Capturing the incongruity in the texts is the critical factor in sarcasm detection. Although several studies have looked at the incongruity of a single text, there is currently a lack of studies on modeling the incongruity of contextual information. Inspired by <em>Multi-Head Attention</em> mechanism from Transformer, we propose a <em>Multi-head Incongruity Aware Attention Network</em>, which concentrates on both target semantic incongruity and contextual semantic incongruity. Specifically, we design a multi-head self-match network to capture target semantic incongruity in a single text. Moreover, a multi-head co-match network is applied to model the contextual semantic incongruity. Furthermore, due to the scarcity of sarcasm data and considering the correlation between tasks of sentiment analysis and sarcasm detection, we pre-train the language model with a great amount of sentiment analysis data, which enhances its ability to capture sentimental features in the text. The experimental results demonstrate that our model achieves state-of-the-art performance on four benchmark datasets, with an accuracy gain of 3.8% on Tweets Ghost, 1.1% on SARC Pol, 1.9% on Ciron and an F1-Score gain of 0.3% on FigLang Twitter.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125702"},"PeriodicalIF":7.5,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662119","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"HORSE-CFR: Hierarchical opponent reasoning for safe exploitation counterfactual regret minimization","authors":"Shijia Wang,&nbsp;Jiao Wang,&nbsp;Bangyan Song","doi":"10.1016/j.eswa.2024.125697","DOIUrl":"10.1016/j.eswa.2024.125697","url":null,"abstract":"<div><div>Opponent modeling-based game decision-making algorithms relax the assumption of rationality, having the potential to achieve higher payoffs than Nash equilibrium strategies. For opponent modeling methods, existing work primarily suffers from incompatibility between computational complexity and robustness, leading to difficulties in achieving high payoff decisions from limited historical interactions in imperfect information games. This paper introduces the HORSE-CFR algorithm, which incorporates Hierarchical Opponent Reasoning (HOR) and Safe Exploitation Counterfactual Regret Minimization (SE-CFR) to enhance decision-making robustness in imperfect information games. HOR combines neural networks with Bayesian theory to accelerate reasoning, improve interpretability, and reduce modeling errors. SE-CFR optimizes the balance between profitability and conservatism, integrating opponent modeling-based strategy adaptation into a constrained linear binary optimization framework. In experiments, HORSE-CFR outperformed Nash equilibrium strategies when playing against various opponents, increasing payoffs by 16.4% in Leduc Hold’em and 36.8% in the Transit game, respectively. It also improved payoffs by more than 9.0% compared to the best-known opponent modeling-based safe adaptive algorithm in both games.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125697"},"PeriodicalIF":7.5,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662120","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"SD-ABM-ISM: An integrated system dynamics and agent-based modeling framework for information security management in complex information systems with multi-actor threat dynamics","authors":"Navid Aftabi ,&nbsp;Nima Moradi ,&nbsp;Fatemeh Mahroo ,&nbsp;Farhad Kianfar","doi":"10.1016/j.eswa.2024.125681","DOIUrl":"10.1016/j.eswa.2024.125681","url":null,"abstract":"<div><div>The increasing complexity and dynamic nature of modern Information Systems (IS) and evolving cybersecurity threats pose significant challenges for organizations in managing information security. Traditional methods often focus on isolated security aspects, failing to capture the intricate interdependencies between internal and external threats, vulnerabilities, and defensive strategies. These limitations necessitate a holistic approach that can comprehensively model and analyze the interactions within IS environments. Motivated to address these research gaps, we developed SD-ABM-ISM, a multi-method framework integrating System Dynamics (SD) and Agent-Based Modeling (ABM). This framework is designed to capture the complex dynamics of IS, incorporating insider and outsider threats and their interactions with defensive measures. SD-ABM-ISM enables an in-depth examination of how various threat actors impact security outcomes and how proactive and reactive investment strategies influence the resilience of the IS. The proposed framework provides a unique approach to understanding multi-actor threat dynamics and their effect on IS over time, facilitating informed decision-making for security investments. The framework offers a robust tool for security decision-makers, enabling organizations to align their security strategies with the evolving threat surface and enhance their resilience against cyberattacks. The detailed simulation and statistical analysis identify the influential elements in the IS over time, highlighting the impact of interactions between insider threats, outsider threats, and the IS itself in an environment characterized by high uncertainty and diverse threat behaviors. The insights from these interactions demonstrate how coordinated threats from multiple actors can amplify vulnerabilities while effective security measures can mitigate these risks. Considering proactive and reactive security investment strategies, SD-ABM-ISM provides a dynamic and cost-effective security investment strategy to protect IS from adversaries with various behaviors.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125681"},"PeriodicalIF":7.5,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662209","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A stable soft sensor based on causal inference and graph convolutional network for batch processes","authors":"Jianlin Wang ,&nbsp;Enguang Sui ,&nbsp;Wen Wang ,&nbsp;Xinjie Zhou ,&nbsp;Zebin Zhang ,&nbsp;Ji Li","doi":"10.1016/j.eswa.2024.125692","DOIUrl":"10.1016/j.eswa.2024.125692","url":null,"abstract":"<div><div>Data-driven soft sensor techniques play a crucial role in process control, which can ensure process safety, and improve product quality by measuring key variables that are challenging to measure in batch processes. Batch processes are characterized by periodic batch production. Insufficient utilization of spatiotemporal information and causal relationships between variables in batch process data limits the accuracy of soft sensors, leading to significant intra-batch and inter-batch errors in the models. Accurate and stable soft sensors in batch processes are in great need. In this work, a stable soft sensor based on causal inference and graph convolutional networks is proposed for batch processes. Specifically, a graph structure learning module based on causal inference is employed in order that the network can learn the causal relationships from both global and local causal effects among process variables. Moreover, a causal graph convolutional network is constructed to capture spatial and temporal information and aggregate causal features for soft sensor modeling. Furthermore, the stable soft sensor model is trained end-to-end using a joint loss function. Experimental results from two batch processes demonstrate the feasibility and effectiveness of stable soft sensor, and the learned causal relationships between variables closely correspond to the fundamental principles of the process.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125692"},"PeriodicalIF":7.5,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662126","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"External information enhancing topic model based on graph neural network","authors":"Jie Song ,&nbsp;Xiaoling Lu ,&nbsp;Jingya Hong ,&nbsp;Feifei Wang","doi":"10.1016/j.eswa.2024.125709","DOIUrl":"10.1016/j.eswa.2024.125709","url":null,"abstract":"<div><div>In the digital age, social media platforms have seen a surge in user-generated content, particularly short-form we-media content. Traditional topic modeling methods often struggle to effectively analyze such content due to their limited generalization ability and interpretability. To address this issue, we propose the Co-occurrence Graph Topic Model (COGTM), a novel approach designed to enhance topic modeling in the context of long-short text co-occurrence scenarios. COGTM leverages the inherent interconnectedness between short and associated long-texts, as well as semantically similar words, within the text corpus. By incorporating these associations into the modeling process, COGTM aims to capture richer semantic information and improve the interpretability of the learned topics. Empirical analysis demonstrates that COGTM outperforms baseline models in various text classification and clustering tasks. By effectively capturing the latent associations between different types of text elements, COGTM offers a promising approach to topic modeling in scenarios involving diverse and interconnected textual data.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125709"},"PeriodicalIF":7.5,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662124","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Numerical computation of Cross nanofluid model using neural network and Adaptive Neuro-Fuzzy Inference system with statistical insights for enhanced flow optimization","authors":"Fuzhang Wang ,&nbsp;Sadique Rehman ,&nbsp;Majid Hussain Shah ,&nbsp;Mohamed Anass El Yamani ,&nbsp;Sohail Farooq ,&nbsp;Aamir Farooq","doi":"10.1016/j.eswa.2024.125721","DOIUrl":"10.1016/j.eswa.2024.125721","url":null,"abstract":"<div><div>In this study, we present a novel integration of numerical methodologies and advanced computational intelligence to elucidate the dynamics of cross nanofluid flow over a Riga plate through a Darcy-Forchheimer porous medium. Brownian motion and thermophoretic phenomena are considered, along with the impacts of activation energy. Slip velocities, convective, and zero-flux boundary conditions are taken into account in the 3D cross nanofluid flow model in the presence of gyrotactic microorganisms. The non-linear PDE model is transformed into a highly non-linear ODE system using von Kármán similarity variables. Taking advantage of Python-derived numerical data as a foundational dataset from the system of non-linear ODEs, we employ neural network algorithms to refine and predict flow behaviors under varying conditions. The research progresses by contrasting these predictions with empirical observations, providing a rigorous validation framework. Furthermore, we incorporate the Adaptive Neuro-Fuzzy Inference System (ANFIS) alongside statistical analyses to examine the impacts of physical parameters, offering unparalleled insight into nanofluid mechanics. This multifaceted approach not only bridges theoretical and practical aspects of fluid dynamics but also proposes a robust model for predicting nanofluid behavior, poised to catalyze advancements in thermal engineering and nanotechnology applications. The precision and adaptability of our methodology underscore its potential as a cornerstone in future fluid dynamics research, inviting scrutiny and discussion from esteemed peers in the field. We have validated our model by finding various sets of error estimations. Furthermore, the velocity profile increases with the enhancement of the Hartmann number or magnetic parameter while decreasing with higher values of the Weissenberg number. The temperature profile decreases with increasing estimates of thermal stratification and the Biot number. The concentration profile amplifies with the Brownian motion parameter while decreasing against the thermophoresis parameter.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125721"},"PeriodicalIF":7.5,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662276","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Study on runoff forecasting and error correction driven by atmosphere–ocean-land dataset","authors":"Xinyu Chang ,&nbsp;Jun Guo ,&nbsp;Yi Liu ,&nbsp;Xiangqian Wei ,&nbsp;Xinying Wang ,&nbsp;Hui Qin","doi":"10.1016/j.eswa.2024.125744","DOIUrl":"10.1016/j.eswa.2024.125744","url":null,"abstract":"<div><div>Accurate runoff forecasting results can not only provide an important basis for flood control scheduling, but also provide scientific support for water resources optimization, which promotes the maximization of the overall benefits of the basin. To further explore the inherent mechanisms of the atmosphere–ocean-land factors driving runoff changes, this study proposes the factors dimension reduction and interpretation framework based on Pearson, eXtreme Gradient Boosting and SHapley Additive exPlanations (P-XGBoost-SHAP). Base on this, the Gaussian Process Regression (GPR), Long Short-Term Memory neural network (LSTM) and Support Vector Machine (SVM) models are used to construct the atmospheric-ocean-land data-driven runoff prediction model. Meanwhile, for the runoff prediction residuals, this paper proposes an error multi-step correction framework based on ensemble empirical mode decomposition and autoregressive model (EEMD-AR). The case study of Lianghekou hydrological station shows that the factor dimension reduction and interpretation framework can greatly reduce the input dimension of the model, and explain the factors globally and locally by using SHAP Value. Compared with the traditional Random Forest (RF) dimension reduction method, it shows higher prediction accuracy. The Nash-Sutcliffe efficiency coefficient (<em>NSE</em>) can be increased to about 0.93, which is 4.91 % and 1.97 % higher than the series–parallel coupling (AR-Parallel) and empirical mode decomposition-autoregressive (EMD-AR) correction methods, respectively. The accuracy of the runoff forecasting prediction is improved while reducing the input dimension of the model.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125744"},"PeriodicalIF":7.5,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662205","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Dynamic group decision-making for enterprise resource planning selection using two-tuples Pythagorean fuzzy MOORA approach","authors":"B.S. Mahapatra ,&nbsp;Debashis Ghosh ,&nbsp;Dragan Pamucar ,&nbsp;G.S. Mahapatra","doi":"10.1016/j.eswa.2024.125675","DOIUrl":"10.1016/j.eswa.2024.125675","url":null,"abstract":"<div><div>Enterprise resource planning (ERP) is an artificial intelligence software that enables organizations to automate and efficiently administer their essential business processes. ERP software assists an organization with time and resource optimization for efficient operations. Choosing an appropriate ERP system for a specific organization has become challenging for decision-makers (DMs) since it involves navigating competing technical, organizational, and financial considerations. A complete, flexible, multi-criteria group decision-making (MCGDM) method based on the Pythagorean fuzzy set (PFS) is proposed in this article to help an organization choose the best option. The PFSs are used to rate both the criteria of an ERP alternative and the quality of the DMs. The collective assessment of the DMs, along with their preferences, is integrated using the Pythagorean fuzzy Einstein-ordered weighted operator (PFEOWO). The method based on the Removal Effects of Criteria (MEREC) is enhanced in a Pythagorean setting to find the Pythagorean weights of each ERP criterion. The Pythagorean weights of the criteria are combined with the DM’s preference using the Pythagorean fuzzy weighted average operator (PFWAO). These combined Pythagorean weights are integrated with the combined assessment of the DMs to produce the weighted Pythagorean decision matrix. Then, the MOORA approach is also enhanced in a PFS setting, and an improved Pythagorean fuzzy MOORA (PF-MOORA) is proposed to solve an MCGDM problem in an uncertain context. Data is generated for three DMs and five ERP packages with twelve characteristics to demonstrate the proposed PF-MOORA. Based on the suggested PF-MOORA method, the fifth and fourth ERP packages are optimal and least favorable, respectively. A sensitivity analysis is conducted by altering the criteria weights to measure the influence of the ERP ranking on the criteria weights. Finally, the proposed PF-MOORA is compared with existing crisp and Pythagorean multi-criteria decision analysis methods to demonstrate its coherence and resilience.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125675"},"PeriodicalIF":7.5,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662206","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-scale structure-guided graph generation for multi-view semi-supervised classification","authors":"Yilin Wu ,&nbsp;Zhaoliang Chen ,&nbsp;Ying Zou ,&nbsp;Shiping Wang ,&nbsp;Wenzhong Guo","doi":"10.1016/j.eswa.2024.125677","DOIUrl":"10.1016/j.eswa.2024.125677","url":null,"abstract":"<div><div>Graph convolutional network has emerged as a focal point in machine learning because of its robust graph processing capability. Most existing graph convolutional network-based approaches are designed for single-view data, yet in many practical scenarios, data is represented through multiple views. Moreover, due to the complexity of multiple views, normal graph generation methods cannot mitigate redundancy to generate a high quality graph. Although the ability of graph convolutional network is undeniable, the quality of graph directly affects its performance. To tackle the aforementioned challenges, this paper proposes a multi-scale graph generation deep learning framework, called multi-scale semi-supervised graph generation based multi-view classification, consisting of two modules: edge sampling and path sampling. The former aims to generate an adjacency graph by selecting edges based on the maximum likelihood among graphs from different views. Meanwhile, the latter seeks to construct an adjacency graph according to the characteristics of paths within the graphs. Finally, the statistical technique is employed to extract commonality and generate a fused graph. Extensive experimental results robustly demonstrate the superior performance of our proposed framework, compared to other state-of-the-art multi-view semi-supervised approaches.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125677"},"PeriodicalIF":7.5,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662188","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}