Cyber threat detection has become an important area of focus in the digital age of today due to the growing spread of fake information and harmful content on social media platforms such as Twitter (now “X”). These cyber threats, often disguised within tweets, pose significant risks to individuals, communities, and even nations, emphasizing the need for effective detection systems. While previous research has explored tweet-based threats, much of the work is limited to specific languages, domains, or locations or relies on single-model approaches, reducing their applicability to diverse real-world scenarios. To address these gaps, our study focuses on multilingual tweet cyber threat detection using a variety of advanced models. The research was conducted in three stages: 1) we collected and labeled tweet datasets in four languages—English, Chinese, Russian, and Arabic—employing both manual and polarity-based labeling methods to ensure high-quality annotations; 2) each dataset was analyzed individually using ML and DL models to assess their performance on distinct languages; and 3) finally, we combined all four datasets into a single multilingual dataset and applied deep learning (DL) and large language model (LLM) architectures to evaluate their efficacy in identifying cyber threats across various languages. Our results show that among machine learning models, random forest (RF) attained the highest performance, and the bidirectional long short-term memory (Bi-LSTM) architecture consistently surpassed other DL and LLM architectures across all datasets. These findings underline the effectiveness of Bi-LSTM in multilingual cyber threat detection.

Compared to traditional models, agentic AI represents a highly valuable target for potential attackers as they possess privileged access to data sources and API tools, which are traditionally not incorporated into classical agents. Unlike a typical software application residing in a Demilitarized Zone (DMZ), agentic LLMs consciously rely on nondeterministic behavior of the AI (only defining a final goal, leaving the path selection to LLM). This characteristic introduces substantial security risk to both operational security and information security. Most common existing defense mechanism rely on detection of malicious intent and preventing it from reaching the LLM agent, thus protecting against jailbreak attacks such as prompt injection. In this paper, we present an alternative approach, LLMZ+, which moves beyond traditional detection-based approaches by implementing prompt whitelisting. Through this method, only contextually appropriate and safe messages are permitted to interact with the agentic LLM. By leveraging the specificity of context, LLMZ+ guarantees that all exchanges between external users and the LLM conform to predefined use cases and operational boundaries. Our approach streamlines the security framework, enhances its long-term resilience, and reduces the resources required for sustaining LLM information security. Our empirical evaluation demonstrates that LLMZ+ provides strong resilience against the most common jailbreak prompts. At the same time, legitimate business communications are not disrupted, and authorized traffic flows seamlessly between users and the agentic LLM. We measure the effectiveness of approach using false positive and false negative rates, both of which can be reduced to 0 in our experimental setting.

Providing contextual and comprehensive medical information tailored to individual patients is critical for enabling effective care in the healthcare domain. However, existing approaches often struggle to deliver personalized responses due to the distributed nature of medical data across multiple sources such as patient records, medical literature, and online resources. To address this challenge, we present MedInsight , a multi-source context augmentation framework that leverages Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) techniques to augment patient-specific information from medical transcripts with trusted knowledge from textbooks and web resources to generate personalized and contextually relevant responses. Our framework consists of three phases: patient context retrieval, medical knowledge retrieval, and response generation. By augmenting patient context with relevant external knowledge, MedInsight generates contextually relevant responses, empowering patients and caregivers with actionable insights. Experiments on the MTSamples dataset validate MedInsight ’s effectiveness in generating contextually appropriate medical responses, using a comprehensive set of metrics including RAGAs, TruLens, ROUGE, and BertScore. Additionally, qualitative evaluations by Subject Matter Experts (SMEs) further confirm the relevance and factual correctness of the generated responses.

Predictive maintenance has been used to optimize system repairs in the industrial, medical, and financial domains. This technique relies on the consistent ability to detect and predict anomalies in critical systems. AI models have been trained to detect system faults, improving predictive maintenance efficiency. Typically there is a lack of fault data to train these models, due to organizations working to keep fault occurrences and down time to a minimum. For newly installed systems, no fault data exists since they have yet to fail. By using diffusion models for synthetic data generation, the complex training datasets for these predictive models can be supplemented with high level synthetic fault data to improve their performance in anomaly detection. By learning the relationship between healthy and faulty data in similar systems, a diffusion model can attempt to apply that relationship to healthy data of a newly installed system that has no fault data. The diffusion model would then be able to generate useful fault data for the new system, and enable predictive models to be trained for predictive maintenance. The following paper demonstrates a system for generating useful, multivariate synthetic data for predictive maintenance, and how it can be applied to systems that have yet to fail.

This paper presents ClinicSum, a novel framework designed to automatically generate clinical summaries from patient-doctor conversations. It utilizes a two-module architecture: a retrieval-based filtering module that extracts Subjective, Objective, Assessment, and Plan (SOAP) information from conversation transcripts, and an inference module powered by fine-tuned Pre-trained Language Models (PLMs), which leverage the extracted SOAP data to generate abstracted clinical summaries. To fine-tune the PLM, we created a training dataset of consisting 1,473 conversations-summaries pair by consolidating two publicly available datasets, FigShare and MTS-Dialog, with ground truth summaries validated by Subject Matter Experts (SMEs). ClinicSum's effectiveness is evaluated through both automatic metrics (e.g., ROUGE, BERTScore) and expert human assessments. Results show that ClinicSum outperforms state-of-the-art PLMs, demonstrating superior precision, recall, and F-1 scores in automatic evaluations and receiving high preference from SMEs in human assessment, making it a robust solution for automated clinical summarization.

For data-constrained, complex and dynamic industrial environments, there is a critical need for transferable and multimodal methodologies to enhance anomaly detection and therefore, prevent costs associated with system failures. Typically, traditional PdM approaches are not transferable or multimodal. This work examines the use of Large Language Models (LLMs) for anomaly detection in complex and dynamic manufacturing systems. The research aims to improve the transferability of anomaly detection models by leveraging Large Language Models (LLMs) and seeks to validate the enhanced effectiveness of the proposed approach in data-sparse industrial applications. The research also seeks to enable more collaborative decision-making between the model and plant operators by allowing for the enriching of input series data with semantics. Additionally, the research aims to address the issue of concept drift in dynamic industrial settings by integrating an adaptability mechanism. The literature review examines the latest developments in LLM time series tasks alongside associated adaptive anomaly detection methods to establish a robust theoretical framework for the proposed architecture. This paper presents a novel model framework (AAD-LLM) that doesn’t require any training or finetuning on the dataset it is applied to and is multimodal. Results suggest that anomaly detection can be converted into a "language" task to deliver effective, context-aware detection in data-constrained industrial applications. This work, therefore, contributes significantly to advancements in anomaly detection methodologies.

This research-to-practice full paper presents Bark-plug v.2, a Large Language Model (LLM)-based chatbot system built using Retrieval Augmented Generation (RAG) pipelines to enhance the user experience and access to information within academic settings. The objective of Barkplug v.2 is to provide information to users about various campus resources, including academic departments, programs, campus facilities, and student resources at a university setting in an interactive fashion. Our system leverages university data as an external data corpus and ingests it into our RAG pipelines for domain-specific question-answering tasks. We evaluate the effectiveness of our system in generating accurate and pertinent responses for Mississippi State University, as a case study, using quantitative measures, employing frameworks such as Retrieval Augmented Generation Assessment (RAGAS). Furthermore, we evaluate the usability of this system via subjective satisfaction surveys using the System Usability Scale (SUS). Our system demonstrates impressive quantitative performance, with a mean RAGAS score of 0.96, and satisfactory user experience, as validated by usability assessments.

In recent years, the rapid growth of IoT devices has posed challenges for traditional cloud computing architectures due to issues related to latency, bandwidth, and real-time response. To overcome those constraints, fog computing architecture emerged as a promising solution, responding to IoT device requests at the edge and forwarding them to the cloud only when necessary. However, some requests still need to be sent to the cloud, making them susceptible to the shortcomings of cloud-based systems. To enhance the fog computing approach, this research proposes an integrated peer-to-peer (P2P) architecture within the fog layer. The proposed model builds upon the 2-layer LDEPTH architecture for P2P networks, wherein current networks are organized hierarchically, with each of the networks having a diameter of one overlay hop. These modest diameters play a crucial role in the new P2P fog model, as they facilitate highly effective data lookup techniques. In this architecture, fog nodes collaborate to handle client requests, enabling them to be fulfilled within the fog layer, thereby reducing the load on the cloud and improving overall efficiency. Moreover, to enhance communication security, we employ ciphertext policy attribute-based encryption (CP-ABE), which introduces a novel secure algorithm. The proposed architecture is further evaluated through simulations using omnet++, with experimental results compared to existing research work based on Distributed Hash Table (DHT) protocols.

Motion planning of an autonomous mobile robot is involved in generating safe, optimal, short, and/or reasonable trajectories in its workspace and finally reaching its final target while avoiding collision with obstacles and escaping traps. This paper presents a new hybrid model to optimize trajectory of the global path of a mobile robot using a graph-based search algorithm associated with an ant colony optimization (ACO) method. Once a graph representing the robot workspace populated with obstacles is modelled by MAKLINK graph theory, Dijkstra algorithm is utilized to seek the sub-optimal collision-free robot trajectory. On the basis of the initial global sub-optimal trajectory generated by Dijkstra algorithm, the motion trajectory of the mobile robot is optimized in Cartesian space through the ACO method. Most importantly, a Bspline curve based smoothing scheme is, in a greater degree, applied to generate safer and smother trajectories with reasonable distance from obstacles. Results of simulation and comparison studies in various sorts of environments are addressed in order to demonstrate the superiority of the proposed hybrid graph-based model.

We introduce and compare two prediction systems on the task of replicating human decisions regarding patient admittance in a typical American emergency department. The data-set used describes the patient trajectories in a 65,000 patient per-year emergency department in the United States. Among the descriptive attributes those of prime importance are the severity of the patient's condition and the time they waited to be admitted from the waiting room to the department proper. A recurrent neural network (RNN) is developed to learn the task of selecting the next patient from the waiting-room/queue to be admitted for treatment which is then compared to a heuristic-based selection algorithm currently used in industry for hospital simulation applications. We demonstrate achievable accuracies of 75.29% and 84.97% using the RNN, depending on the type of the data preprocessing used. These accuracies are only potentially and theoretically achievable, respectively. The former's validity hinges on whether certain "anomalous cases" are outliers or not, the second is achieved with the assumed existence of a method for labeling these same cases as anomalous as part of the RNN's input, which may or may not be achievable, pending further consultation with industry experts. Our conclusions hinge on whether or not such cases are outliers though in either case a more sophisticated data-set is desired. If they are not outliers then a more detailed data-set is likely necessary to apply machine learning, or at least our methods, meaningfully to this prediction problem for use in simulated, or real world, hospitals.

Machine Learning (ML) aims at designing models that learn from previous experience, without being explicitly formulated. Applications of machine learning are inexhaustible, including recognizing patterns, predicting future trends and making decisions, and they are capable of handling sizable quantities of multi-dimensional data in the form of large vectors and tensors. To perform these operations on classical computers, however, requires vast time and computational resources. Unlike the classical computers that rely on computations using binary bits, Quantum Computers (QC) benefit from qubits which can hold combinations of 0 and 1 at the same time via superposition and entanglement. This makes QCs powerful at handling and post processing large tensors, making them a prime target for implementing ML algorithms. While several models used for ML on QCs are based on concepts from their classical computing counterparts, utilization of the QC's potential has made them the superior of the two. This paper presents an overview of the current state of knowledge in application of ML on QC, and evaluates the speed up, and complexity advantages of using quantum machines.

High Performance Computing (HPC) is the ability to process data and perform complex calculations at extremely high speeds. Current HPC platforms can achieve calculations on the order of quadrillions of calculations per second with quintillions on the horizon. The past three decades witnessed a vast increase in the use of HPC across different scientific, engineering and business communities, for example, sequencing the genome, predicting climate changes, designing modern aerodynamics, or establishing customer preferences. Although HPC has been well incorporated into science curricula such as bioinformatics, the same cannot be said for most computing programs. This working group will explore how HPC can make inroads into computer science education, from the undergraduate to postgraduate levels. The group will address research questions designed to investigate topics such as identifying and handling barriers that inhibit the adoption of HPC in educational environments, how to incorporate HPC into various curricula, and how HPC can be leveraged to enhance applied critical thinking and problem solving skills. Four deliverables include: (1) a catalog of core HPC educational concepts, (2) HPC curricula for contemporary computing needs, such as in artificial intelligence, cyberanalytics, data science and engineering, or internet of things, (3) possible infrastructures for implementing HPC coursework, and (4) HPC-related feedback to the CC2020 project.

Trading cryptocurrencies (digital currencies) are currently performed by applying methods similar to what is applied to the stock market or commodities; however, these algorithms are not necessarily well-suited for predicting cryptocurrency prices. Unlike stock exchanges, which shut down for several hours or days at a time, digital currency prediction and trading seem to be of a more consistent and predictable nature. In this work, we benefit from sentiment analysis of tweets using both an existing sentiment analysis package and a manually tailored “objective analysis,” to calculate one impact value for each analysis every 15[Formula: see text]min. We then select the most appropriate training method by applying evolutionary techniques and discover the best subset of the generated features to include, as well as other parameters. One of the unique contributions of this work is the analysis of both English and Japanese tweets with a tailored “objective analysis” tool. This resulted in implementation of predictors which yielded 28% to 122% profit in a four-week simulation, much more than simply holding a digital currency for the same period of time.

Sepsis is the leading cause of death in ICUs and a very costly medical phenomena. The earlier it is predicted, the less inpatient mortality and the less the length of ICU stay, thus a major cut in medical expenses. Although the current deep learning models are able to make predictions about the possibility of sepsis in the ICU, they still lack the ability to reveal the major factors that lead to the outcomes of the predictions. In this paper, we have explored the use of an attention-based model in prediction of sepsis which provides more details on the amount of contribution of each of the medical measurements to the final prediction. This would help health care providers to improve their procedures to reduce sepsis related mortality rate.

In this work, we have considered a new hierarchical non-DHT based architecture for Peer-to-Peer (P2P) networks in which at each level of the hierarchy existing networks are all structured and each such network has the diameter of l overlay hop. Such low diameters have immense importance in designing very efficient data lookup algorithms. The mathematical model based on the Chinese Remainder Theorem has been used to define the neighborhood relations among peers to obtain the above-mentioned diameters. In this work, we present very efficient intra-group as well as inter-group data look up algorithms with O (1) time complexity exploiting the above mentioned low diameter features of the P2P system. Besides, we present very efficient algorithms related to peers joining with new and existing resource types.

Facebook is the largest and the most popular online social network application that records large amount of users’ behavior expressed in various activities such as Facebook likes, status updates, posts, comments, photos, tags and shares. One of the major attractions of such a dataset relates to the predictability of the individuals’ psychological traits from their digital footprints. Such predictions help researchers and service providers to improve personalized offering of products and services. The goal of this work is to investigate the predictability of Facebook users’ personality traits, measured by BIG5 test as a function of their digital records of behavior such as Facebook likes. This research is based on a dataset of 92,255 users who provided their Facebook likes and the results of their BIG5 personality test. As the Facebook likes data includes 600 attributes, the proposed model uses LASSO algorithm to select the best features and to predict Facebook users’ BIG5 personality traits. The best accuracy level of these predictions is achieved for Openness and Extraversion, the lowest accuracy level is obtained for Agreeableness while the accuracy levels of Conscientiousness and Neuroticism are in the middle.

Multicasting can be done in two different ways: source based tree approach and shared tree approach. Shared tree approach is preferred over source-based tree approach because in the later construction of minimum cost tree per source is needed unlike a single shared tree in the former approach. However, in shared tree approach a single core needs to handle the entire traffic load resulting in degraded multicast performance. Besides, it also suffers from `single point failure'. In this paper, novel and efficient schemes for load shared multicore multicasting are presented. Multiple cores are selected statically, that is, independent of any existing multicast groups and also the selection process is independent of any underlying unicast protocol. Some of the selected cores can be used for fault-tolerant purpose also to guard against any possible core failures.

Reverse Path Forwarding based multicasting used in Distance Vector Multicast Routing Protocol (DVMRP) has the major drawback of generating large number of duplicate and prune packets. Use of the concept of pseudo diameter reduces the number of duplicate packets. The reason is that pseudo diameter helps in the reduction of the scope of travel of a packet. In this work a new approach of multicasting with security has been proposed. The concept of pseudo diameter has been extended to further reduce the scope of travel of a packet. Use of this new concept known as super pseudo diameter while multicasting results in significant reduction of the number of prune packets generated; hence, it guarantees even better utilization of network bandwidth.

In this paper, we have considered a recently reported non-DHT based structured P2P system. The architecture is based on Linear Diophantine Equation (LDE) and it is an interest-based system; it offers very efficient data lookup. However, the architecture is restricted in that a peer cannot possess more than one distinct resource type. This may reduce the scope of its application. In this paper, we have extended the work by considering a generalization of the architecture, that is, a peer can possess multiple distinct resource types. We have proposed an efficient data lookup algorithm with time complexity bounded by (2+r/2); r is the number of distinct resource types. We have discussed about an alternative lookup scheme that needs constant number of hops and constant number of message exchanges. Besides, churn handling and ring maintenance have been shown to be very efficient.

Sentiment Analysis tools, developed for analyzing social media text or product reviews, work poorly on a Software Engineering (SE) dataset. Since prior studies have found developers expressing sentiments during various SE activities, there is a need for a customized sentiment analysis tool for the SE domain. On this goal, we manually labeled 2000 review comments to build a training dataset and used our dataset to evaluate seven popular sentiment analysis tools. The poor performances of the existing sentiment analysis tools motivated us to build SentiCR, a sentiment analysis tool especially designed for code review comments. We evaluated SentiCR using one hundred 10-fold cross-validations of eight supervised learning algorithms. We found a model, trained using the Gradient Boosting Tree (GBT) algorithm, providing the highest mean accuracy (83%), the highest mean precision (67.8%), and the highest mean recall (58.4%) in identifying negative review comments.

Bayesian decision methods have been widely used in different areas for estimating, predicting, and offering decision supports. On the other hand, computing with words (CW) and approximate reasoning have effectively been used to deal with imprecise measurements and inexact information. This work is an attempt to benefit from both approaches and to combine them to model the expert knowledge in natural language and apply it to Bayesian decision support systems. The proposed method employs z-numbers to represent the knowledge in natural language and its certainty (reliability). To evaluate applicability of the proposed method, an example is provided as a proof of concept.

Surveys are used by hospitals to evaluate patient satisfaction and to improve operation. Collected satisfaction data is usually represented to the hospital administration using statistical charts and graphs. Although this statistical data and visualization is helpful, but because of the size and dimension of the dataset, it is very difficult if not impossible, to identify important factors that could be evaluated and improved for better patient satisfaction. This work presents an unsupervised data-driven methodology that discovers the specific issues reflected by the dataset from the patients' point of view. The goal of the introduced exploratory data analysis methodology is to determine hidden patterns in the dataset and to identify the main causes of dissatisfaction. To this end, two layers of data analysis is performed. In the first layer, the analysis is only performed on the satisfaction questions. The analysis consists of handling the high dimensionality using self-organizing maps, grouping similar patients using clustering methods and labeling each cluster according to their salient features. In the second layer, demographic data of patients of each cluster is fed to the same analysis process. Putting the salient features of a cluster and its sub-clusters together, one can extract correlations. The correlations are validated using multiple statistical methods applied to the dataset. In a following work, the correlations extracted using this methodology will be ranked and converted to recommendations that can be used by healthcare providers as well as patients.

This study aims to optimize patient flow in emergency departments (ED) while minimizing associated costs. In order to compare the effects of the optimization, a simulation model for emergency departments has been implemented using District Event Simulation (DES) and queuing theory, while for the optimization, Genetic Algorithm is used to find the best arrangements. Principally, a discrete event based, multi-class, multi-server queuing network is designed considering the emergency department as a set of stages each associated with a queue of patients waiting to be served. Each stage has multiple service providers such as Nurses, Doctors or other staff. We also classified patients passing through the stages according to their acuity level and personal characteristics. Then, a function is defined to measure the ED performance in respect to the calculated wait times and the cost. Finally, a customized genetic algorithm was developed to discover the best performance which reflects the best allocation of service providers to the different stages of the emergency department.

In this paper, we present a new approach to designing a scalable, hierarchical overlay P2P system which provides highly efficient data lookup operations as well fault tolerance. Instead of using a distributed hash table (DHT) based approach, we propose the use of Linear Diophantine Equation (LDE) as the mathematical basis to realize a hierarchical P2P architecture. LDE provides a significantly lighter weight mechanism to create and maintain the overlaid P2P structure as compared to DHT based schemes. To the best of our knowledge, ours is the first work that proposes the use of LDEs in designing P2P topologies. In our proposed hierarchical P2P architecture, the number of hops required to search for a resource is independent of the number of nodes/peers in the network n and is instead bounded by (1 + r/2), r being the number of distinct resource types. For most practical purposes, r is significantly smaller than n and hence our proposed architecture provides a highly efficient resource look-up procedure. We have also presented algorithms for handling of new resources and peers joining and leaving the P2P network and fault tolerance in the event of peers crashing or leaving.

Some of the attractive features of using Reverse Path Forwarding (RPF) method for broadcasting include reasonably low amount of memory necessary for storing the data structures in a router required by the protocol and ease of its implementation as it does not look for shortest path delivery. However, it generates quite a large number of duplicate packets resulting in poor bandwidth utilization. In this paper we have introduced a novel hierarchical pruning mechanism to improve bandwidth utilization for broadcasting using the RPF method. Experimental results support our theoretical findings of achieving very high reduction in the number of the duplicate packets generated.

The generalized constraint language (GCL), introduced by Zadeh, serves as a basis for computing with words (CW). It provides an agenda to express the imprecise and fuzzy information embedded in natural language and allows reasoning with perceptions. Despite its fundamental role, the definition of GCL has remained informal since its introduction by Zadeh, and to our knowledge, no attempt has been made to formulate a rigorous theoretical framework for GCL. Such formalization is necessary for further theoretical and practical advancement of CW for two important reasons. First, it provides the underlying infrastructure for the development of useful inference patterns based on sound theories. Second, it determines the scope of GCL and hence facilitates the translation of natural language expressions into GCL. This paper is an attempt to step in this direction by providing a formal syntax together with a compositional semantics for GCL. A soundness theorem is defined, and Zadeh's deduction rules are proved to be valid in the defined semantics. Furthermore, a discussion is provided on how the proposed language may be used in practice.

In this paper, we have proposed a simple and efficient approach for recovery in distributed computing environment. The check pointing scheme used in the work ensures that after the system recovers from failures, all processes can restart from their respective recent checkpoints; thus avoiding any domino effect. That is, the recent check points always form a consistent recovery line. The recovery scheme deals with both orphan and lost messages. Therefore, correctness of computation of the underlying application program is guaranteed.

This paper paves the path for the design and implementation of a practical automated reasoning mechanism for computing with words. The presented methodology aims at filling the gap between the raw theory and the implementation of CW paradigm. As part of this methodology, a portion of a formal grammar for Generalized Constraint Language (GCL) as well as a systematic approach for applying the deduction rules are introduced.

Covert channel attacks utilise shared resources to indirectly transmit sensitive information to unauthorised parties. Current operating systems (e.g. SELinux) rely on tagging the filesystem with security labels and enforcing security policies at the time of access to a file or resource. However, such mechanisms do not provide strong protection against information laundering via covert channels. Colored Linux, an extension to SELinux, utilises watermarking algorithms to ‘colour’ the contents of each file with their respective security classification, or context, to enhance resistance to information laundering attacks. In this study, the authors propose a mobile agent-based approach to automate the process of detecting and colouring receptive hosts' filesystems and monitoring the coloured filesystem for instances of potential information leakage. Implementation details and execution results are included to illustrate the merits of the proposed approach. The authors have also evaluated the performance of their agent-based system over a single host as well as a local network of machines. Finally, using formal method techniques, the authors have proved correctness properties about the agent-based approach and identified and corrected a flaw in their initial implementation.

Model-based intrusion detection and knowledge discovery are combined to cluster and classify P2P botnet traffic and other malignant network activity by using a self-organizing map (som) self-trained on denied Internet firewall log entries. The SOM analyzed new firewall log entries in a case study to classify similar network activity, and discovered previously unknown local P2P bot traffic and other security issues.

We propose a model for using firewall log entries of denied inbound Internet traffic for indirect discovery of local IP addresses that have security problems. This method is used successfully to discover two computers on the network of Southern Illinois University which were infected with malicious feral software, as well as two more IP addresses on the university network with other security problems.

Adaptive Web systems utilize user models (or group models) to represent essential information about an individual user (or a group) so that these systems can adapt their behavior to the goals, tasks, interests, and other features of individual users

Multicriteria decision making (MCDM) methods can be powerful aids for evaluating patients' medical information in medical diagnostic systems. Technique ordered preference by similarity to the ideal solution (TOPSIS) is one of the more widely used MCDM methods in decision support systems. For the purpose of this work, the TOPSIS method is modified into a more suitable form and used for the implementation of a web-based medical diagnostic system. In our modified TOPSIS method, we have utilized fuzzy logic so that users can more accurately describe their symptoms. The data given to the modified TOPSIS method are often massive in proportions and may take a considerable amount of time to generate a ranking of alternatives. TOPSIS lends itself to parallel computation because it is virtually a combination of matrix computations. Therefore, computer parallelism is implemented so that a large amount of input data can be handled simultaneously, hence decreasing overall execution time. In addition, to make our MCDM system more accessible, we have designed our system to be web based. The web-based medical diagnosis system includes a dynamically generated web-based user interface, while the parallel implementation of the modified TOPSIS component, in conjunction with the Common Gateway Interface, forms the back end of the system. © 2007 Wiley Periodicals, Inc. Int J Int Syst 22: 1083–1099, 2007.

Clustering algorithms are widely used methods for organizing data into useful information. The Competitive Agglomeration for Relational Data (CARD) Algorithm is one such clustering algorithm that is designed to organize user sessions into profiles, where each profile would highlight a particular type of user. The CARD algorithm is a viable candidate for Web clustering; however, it does have limitations such as an extended execution time. In addition, the methods that prepare the input data for the CARD algorithm's use employs concepts which seem to be incomplete. These limitations of the CARD algorithm are explored and modifications are introduced to yield a more practical and efficient algorithm.

The current search engine technologies mostly use a keyword based searching mechanism, which does not have any deductive abilities. There is an urgent need for a more intelligent question-answering system that will provide a more intuitive, natural language interface, and more accurate and direct search results. The introduction of computing with words (CwW) provides a new theoretical base for developing frameworks with support for dealing with information in natural language. This paper proposes a domain specific question-answering system based on fuzzy expert systems using CwW. In order to perform the translation of natural language based information into a standard format for use with CwW, probabilistic context-free grammar is used

For many years, spatial querying has been of interest for the researchers in the GIS community. Any successful implementation and long-term viability of the GIS technology depends on the issue of accuracy of spatial queries. In order to improve the accuracy and quality of spatial querying, the problems associated with the areas of fuzziness and uncertainty need to be addressed. There has been a strong demand to provide approaches that deal with inaccuracy and uncertainty in GIS. In this paper, we develop an approach that can perform fuzzy spatial querying under uncertainty. An inexact inferencing strategy for objects with determined and indeterminate boundaries is investigates using type-2 fuzzy set theory.

Intelligent mobile agent technology is one of the most promising of the newer software paradigms for providing solutions to complex, distributed computing problems. Agent properties of autonomy, intelligence and mobility provide a powerful platform for implementations that can utilize techniques involving collaborative problem solving and adaptive behavior. Although the technological tools and capabilities have advanced to this point, research into formal models and extensions to support representations of this new computing paradigm has not been kept pace. Specifically, we find that current formal processing models are lacking in representation abilities for: (1) intelligence capabilities, (2) team-based problem-solving approaches, and (3) mobility. In this paper, we present an extension of /spl pi/-calculus that addresses the first of these deficiencies, the representation of intelligence.