Obstructive Sleep Apnea (OSA) management struggles from static retrospective adherence models which build their assessment on rigid compliance benchmarks. This paper describes a modular AI-based framework for CPAP (Continuous Positive Airway Pressure) personalization, which combines real-time digital biomarkers and dynamic therapystage clustering. Patients are stratified into four progressive stages, new, adherent, attempting and non-adherent, using unsupervised learning on the CPAP usage data, clinical history, and behavioral surveys. The system dynamically adjusts biomarker tracking such as HRV, ODI and RVO, produces weekly risk scores and triggers phase specific interventions such as alerts, education modules and physician reviewed actions. Unlike previous models, the framework continually improves clusters and provides interventions depending on physiological feedback and engagement history. Preliminary timescale review shows its ability of identifying dropout threat and adaptation of interventions in response to real time surveillance. The proposed system enables scalable, explainable, and medically integrated for biomarker-driven real-time CPAP personalization methods throughout the OSA care pathway.

International audience

Abstract Data interoperability implies data exchanges among intra and inter enterprises collaborating with information systems (IS). The multiplicity of these exchanges and the increasing number of data exchanged generates complexity and brings out the needs for control to avoid dysfunctions with a negative impact on the overall performance of the systems. Indeed, actually, interoperability has become a necessary performance lever that thus requires particular attention. Being at a low level in the enterprise interoperability concerns, data interoperability is mainly automated, which leads us to question: Is it possible to evaluate data interoperability performance and security based on inspection and analysis of ongoing data exchanges? We therefore endeavored to answer this question by establishing monitoring and analysis systems. In this paper, we present a research work which addresses services provided by a messaging‐based communication system. In order to collect information on Information System interactions allowing one to assess their level of data interoperability, we propose a messaging metamodel that aggregates the collected information. It provides a single point of control and enables one to determine indicators of potential interoperability problems. The approach is validated on two case studies. An industrial case study of interactions among existing systems is presented to showcase the feasibility and interest of the approach. It is proposed on top of RabbitMQ and allows our partner to identify some issues in the studied information system. The second case study shows that the approach can integrate other protocols, by reading MQTT messages. The approach is implemented using Moose, a software analysis platform.

International audience

International audience

Obstructive Sleep Apnea (OSA) is a potentially severe sleep disorder that leads to different pathology. The goal treatment to OSA is the Positive Airway Pressure (PAP) therapy. Nevertheless, this therapy has one of the lowest compliance levels when compared to the other 17 therapies. For the last two decades, trials were carried out to improve this compliance level and understand factors impacting compliance, but there were no conclusive results. In this paper, we propose a framework for modeling multiple patient profiles at a different moment in the PAP therapy. This approach in PAP therapy takes into consideration multiple factors and the interactions between the factors at a specific moment in the therapy to understand and tackle the compliance problem. The data pre-processing is implemented in Python to extract the factors from the raw data. The processing and the core features of the framework are implemented in R. Six different patient profile was identified based on the event recorded between 3 days and 15 days after the installation of the PAP device at the patient home.

International audience

The role of Information Technology (IT) is critical in supporting business processes. This role is even more critical when organizations operate in a dynamic, uncertain and extremely competitive environment. The growing role of IT in organizations imposes the alignment of their use with business processes requirements and company strategy. The necessity of Business/IT alignment is a key to improve business performance. This paper proposes a methodology for analyzing and evaluating the quality of business process execution. The aim of this methodology is to assess the business process during its lifecycle and to propose some solutions to restore the alignment between IT infrastructure and Business Process in a collaborative context. Our approach is composed of 5 models, (1) analysis reference model (risk, maturity...) for analyzing the component execution traces in business processes, (2) ontological model to ensure the description of processes between the applicative level (execution), functional level and business level, (3) interpretation model (butom-up) for aggregating the performance of processes at business level, (4) assessment business process model based on Key Indicators Performance and (5) aggregation model to validate the final performance for collaborative business process.

The structures of product design, process development, manufacturing, sales, product utilization, after sale service and product retirement are becoming more and more complex due to the underlying reasons of newly introduced rules as well as constraints of the respective stakeholders. On the other hand, these complexities are also appending up the organizational assets. The configuration of the state of art and introduction of an architecture which can convert this organizational asset into an action such as "to introduce an added value" in the product life cycle is an interesting but challenging task. We in this study have discussed the issues related to enterprise collaboration for the purpose of assessment of an enterprise's capability. Moreover, we have also proposed the solution for how we can utilize the structural knowledge of ontological modeling to address this issue along with the emerging technological advances. It's flexible as well as useful to utilize the state of the art digital preservation model to store, reuse and access the ontologies.

Ontology mapping and merging systems play a vital role that aim at promoting automatic interoperability among different heterogeneous systems, agents, web services or groups in open environments such as Semantic Web. These systems help ontologists to resolve different types of conflicts among local ontologies to produce global merged ontology. This paper provides three contributions to the study and design of ontology merging systems that provides complete, consistent and coherent merged global ontology. First, we analyze that one of the important merge requirements is ignored yet by state-of-the-art ontology mapping and merging systems, i.e., Disjoint-knowledge Preservation between concepts. Second, we introduce another type of semantic conflict, which needs attention for consistent and coherent merged ontology, i.e., Alignment Conflict among disjoint relations. Third, we present an overview of our semantic-based ontology merger, DKP-OM, as a solution for the generation of global merged ontology that is consistent, coherent and complete with respect to local ontologies. We conclude that disjoint knowledge analysis for ontology merging is very much helpful for the detection of inconsistent initial mappings that originate from concept name or instance matching strategies, reduce search space for concept matching, and promote consistent computation by exploiting reliable logical inference on facts by axiomatization.