A reflective analysis on the experience of virtual environment (VE) design is presented focusing on the human–computer interaction (HCI) challenges presented by virtual reality (VR). HCI design guidelines were applied to development of two VRs, one in marine archaeology and the other in situation awareness simulation experiments. The impact of methods and HCI knowledge on the VR design process is analyzed, leading to proposals for presenting HCI and cognitive knowledge in the context of design trade-offs in the choice of VR design techniques. Problems reconciling VE and standard Graphical User Interface (GUI) design components are investigated. A trade-off framework for design options set against criteria for usability, efficient operation, realism, and presence is proposed. HCI-VR design advice and proposals for further research aimed towards improving human factor-related design in VEs are discussed.
6 June 2005 in IEEE Transactions on Software Engineering
This paper describes a method and a tool for validating nonfunctional requirements in complex socio-technical systems. The system requirements analyzer (SRA) tool validates system reliability and operational performance requirements using scenario-based testing. Scenarios are transformed into sequences of task steps and the reliability of human agents performing tasks with computerized technology is assessed using Bayesian belief network (BN) models. The tool tests system performance within an envelope of environmental variations and reports the number of tests that pass a benchmark threshold. The tool diagnoses problematic areas in scenarios representing pathways through system models, assists in the identification of their causes, and supports comparison of alternative requirements specifications and system designs. It is suitable for testing socio-technical systems where operational scenarios are sequential and deterministic, in domains where designs are incrementally modified so set up costs of the BNs can be defrayed over multiple tests.
Revisit intention is a key indicator of business performance, studied in many fields including hospitality. This work employs big data analytics to investigate revisit intention patterns from tourists’ electronic word of mouth (eWOM) using text classification, negation detection, and topic modelling. The method is applied on publicly available hotel reviews that are labelled automatically based on consumers’ intention to revisit a hotel or not. Topics discussed in revisit-annotated reviews are automatically extracted and used as features during the training of two Extreme Gradient Boosting models (XGBoost), one for each of two hotel categories (2/3 and 4/5 stars). The emerging patterns from the trained XGBoost models are identified using an explainable machine learning technique, namely SHAP (SHapley Additive exPlanations). Results show how topics discussed by tourists in reviews relate with revisit/non revisit intention. The proposed method can help hoteliers make more informed decisions on how to improve their services and thus increase customer revisit occurrences.
1 April 2008 in Reliability Engineering & System Safety
The development of universal methodologies for the accurate, efficient, and timely prediction of traffic accident location and severity constitutes a crucial endeavour. In this piece of research, the best combinations of salient accident-related parameters and accurate accident severity prediction models are determined for the 2005 accident dataset brought together by the Republic of Cyprus Police. The optimal methodology involves: (a) information mining in the form of feature selection of the accident parameters that maximise prediction accuracy (implemented via scatter search), followed by feature extraction (implemented via principal component analysis) and selection of the minimal number of components that contain the salient information of the original parameters, which combined bring about an overall 74.42% reduction in the dataset dimensionality; (b) accident severity prediction via probabilistic neural networks and random forests, both of which independently accomplish over 96% correct prediction and a balanced proportion of under- and over-estimations of accident severity. An explanation of the superiority of the optimal combinations of parameters and models is given, as is a comparison with existing accident classification/prediction approaches