Local Data Repositories Research Articles

Survey on Federated Learning for Intrusion Detection System: Concept, Architectures, Aggregation Strategies, Challenges, and Future Directions

Intrusion Detection Systems (IDS) are essential for securing computer networks by identifying and mitigating potential threats. However, traditional IDS face challenges related to scalability, privacy, and computational demands as network data complexity increases. Federated Learning (FL) has emerged as a promising solution, enabling collaborative model training on decentralized data sources while preserving data privacy. Each participant retains local data repositories, ensuring data sovereignty and precluding data sharing. Leveraging the FL framework, participants locally train machine learning models on their respective datasets, subsequently transmitting model updates to a central server for aggregation. The central server then disseminates the aggregated model updates to individual participants, collectively striving to bolster intrusion detection capabilities. This article presents a comprehensive survey of FL applications in IDS, covering core concepts, architectural approaches, and aggregation strategies. We evaluate the strengths and limitations of various FL methodologies for IDS, addressing privacy and security concerns and exploring privacy-preserving techniques and security protocols. Our examination of aggregation strategies within the FL framework for IDS aims to highlight their effectiveness, limitations, and potential enhancements.

State-of-the-Art Fast Healthcare Interoperability Resources (FHIR)-Based Data Model and Structure Implementations: Systematic Scoping Review.

Data models are crucial for clinical research as they enable researchers to fully use the vast amount of clinical data stored in medical systems. Standardized data and well-defined relationships between data points are necessary to guarantee semantic interoperability. Using the Fast Healthcare Interoperability Resources (FHIR) standard for clinical data representation would be a practical methodology to enhance and accelerate interoperability and data availability for research. This research aims to provide a comprehensive overview of the state-of-the-art and current landscape in FHIR-based data models and structures. In addition, we intend to identify and discuss the tools, resources, limitations, and other critical aspects mentioned in the selected research papers. To ensure the extraction of reliable results, we followed the instructions of the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) checklist. We analyzed the indexed articles in PubMed, Scopus, Web of Science, IEEE Xplore, the ACM Digital Library, and Google Scholar. After identifying, extracting, and assessing the quality and relevance of the articles, we synthesized the extracted data to identify common patterns, themes, and variations in the use of FHIR-based data models and structures across different studies. On the basis of the reviewed articles, we could identify 2 main themes: dynamic (pipeline-based) and static data models. The articles were also categorized into health care use cases, including chronic diseases, COVID-19 and infectious diseases, cancer research, acute or intensive care, random and general medical notes, and other conditions. Furthermore, we summarized the important or common tools and approaches of the selected papers. These items included FHIR-based tools and frameworks, machine learning approaches, and data storage and security. The most common resource was "Observation" followed by "Condition" and "Patient." The limitations and challenges of developing data models were categorized based on the issues of data integration, interoperability, standardization, performance, and scalability or generalizability. FHIR serves as a highly promising interoperability standard for developing real-world health care apps. The implementation of FHIR modeling for electronic health record data facilitates the integration, transmission, and analysis of data while also advancing translational research and phenotyping. Generally, FHIR-based exports of local data repositories improve data interoperability for systems and data warehouses across different settings. However, ongoing efforts to address existing limitations and challenges are essential for the successful implementation and integration of FHIR data models.

Machine Learning Based Prediction of Postoperative Pancreatic Fistula Following Pancreaticoduodenectomy.

To develop a novel machine learning (ML) model to predict clinically relevant postoperative pancreatic fistula (CR-POPF) following pancreaticoduodenectomy (PD). Accurate prognostication of CR-POPF may allow for risk stratification and adaptive treatment strategies for potential PD candidates. However, antecedent models, such as the modified Fistula Risk Score (mFRS), are limited by poor discrimination and calibration. All records entailing PD within the 2014-2018 ACS NSQIP were identified. Additionally, patients undergoing PD at our institution between 2013 and 2021 were queried from our local data repository. An eXtreme Gradient Boosting (XGBoost) model was developed to estimate the risk of CR-POPF using data from the ACS NSQIP and evaluated using institutional data. Model discrimination was estimated using the area under the receiver operating characteristic (AUROC) and precision recall curve (AUPRC). Overall, 12,281 and 445 patients undergoing PD were identified within the 2014-2018 ACS NSQIP and our institutional registry, respectively. Application of the XGBoost and mFRS scores to the internal validation dataset revealed that the former model had significantly greater AUROC (0.72 vs. 0.68, P<0.001) and AUPRC (0.22 vs. 0.18, P<0.001). Within the external validation dataset, the XGBoost model remained superior to the mFRS with an AUROC of 0.79 (95% CI 0.74-0.84) versus 0.75 (95% CI 0.70-0.80, P<0.001). In addition, AUPRC was higher for the XGBoost model, compared to the mFRS. Our novel ML model consistently outperformed the previously validated mFRS within internal and external validation cohorts, thereby demonstrating its generalizability and utility for enhancing prediction of CR-POPF.

Evaluation of patient STress level caused by radiological Investigations in early Postoperative phase After CRANIOtomy (IPAST-CRANIO): protocol of a Swiss prospective cohort study

IntroductionPostoperative imaging after neurosurgical interventions is usually performed in the first 72 hours after surgery to provide an accurate assessment of postoperative resection status. Patient frequently report that early postoperative...

Heterogeneous Fault Prediction Using Feature Selection and Supervised Learning Algorithms

Software Fault Prediction (SFP) is the most persuasive research area of software engineering. Software Fault Prediction which is carried out within the same software project is known as With-In Fault Prediction. However, local data repositories are not enough to build the model of With-in software Fault prediction. The idea of cross-project fault prediction (CPFP) has been suggested in recent years, which aims to construct a prediction model on one project, and use that model to predict the other project. However, CPFP requires that both the training and testing datasets use the same set of metrics. As a consequence, traditional CPFP approaches are challenging to implement through projects with diverse metric sets. The specific case of CPFP is Heterogeneous Fault Prediction (HFP), which allows the program to predict faults among projects with diverse metrics. The proposed framework aims to achieve an HFP model by implementing Feature Selection on both the source and target datasets to build an efficient prediction model using supervised machine learning techniques. Our approach is applied on two open-source projects, Linux and MySQL, and prediction is evaluated based on Area Under Curve (AUC) performance measure. The key results of the proposed approach are as follows: It significantly gives better results of prediction performance for heterogeneous projects as compared with cross projects. Also, it demonstrates that feature selection with feature mapping has a significant effect on HFP models. Non-parametric statistical analyses, such as the Friedman and Nemenyi Post-hoc Tests, are applied, demonstrating that Logistic Regression performed significantly better than other supervised learning algorithms in HFP models.

PF649 ALPINA: REAL WORLD ANALYSIS OF 1ST LINE RVD TREATMENT IN TRANSPLANT ELIGIBLE AND TRANSPLANT‐NON‐ELIGIBLE MM PATIENTS WITH A FOCUS ON TOLERABILITY AND EFFICACY

Background:The triplet combination of Lenalidomide, Bortezomib, Dexamethasone (RVd) has shown high activity in Myeloma first line treatment in both transplant‐eligible (TE) and transplant non‐eligible (TNE) patients in clinical trials. In Multiple Myeloma (MM), the achievement and duration of complete response (CR) or at least very good partial response (VGPR) has been shown to be of major importance.Aims:As this is particularly relevant in high dose therapy followed by autologous stem cell transplantation (ASCT), in Europe this regimen is not approved yet. In order to analyse the outcome of RVd‐patients in Autria and Germany in a clinical routine setting with a focus on tolerability and efficacy, the ALPINA project was initiated.Methods:Patients receiving RVd were identified using the Austrian Myeloma Registry (AMR) and local data repositories. A minimum of four completed cycles of RVd were required for TE patients, while at least 6 completed cycles of RVd were necessary for TNE patients to qualify for the analysis, making the cohorts more comparable to published trial populations (e.g. SWOG S0777, IFM 2009, GEM 2012). Besides outcome variables (e.g. best response) safety, tolerability, side effects (e.g. neuropathy) and the real‐world (RW) use of regimen modifications were analysed using descriptive statistics.Results:Out of 141 patients, 87 patients met the inclusion criteria [TE (n = 59, 69 % cohort 1), TNE (n = 28, 31% cohort 2)]. Among the patients not included (n = 54), 17 patients discontinued therapy before reaching the specified number of cycles, while 28 patients are still on RVd treatment (but are too early to analyse).Baseline characteristics of the ALPINA cohorts were: 52 male and 35 female MM patients, diagnosed at a median age of 57 (range 43 ‐ 66) in cohort 1 and 69 (range 48 ‐ 77) years in cohort 2. 59 (69 %) MM patients received an ASCT, and 28 (31 %) patients were TNE. 59 (69 %) MM patients received an ASCT, and 28 (31 %) patients were TNE. Within cohort 1, 40 (68 %) patients received a single ASCT, while 19 (32 %) patients received a tandem ASCT. The median number of applied and completed RVd cycles was 4 in cohort 1, compared to 8 cycles in cohort 2.Bortezomib was given subcutaneously in all patients. RVd was applied uniformly 14d/21d in cohort 1 [(V 1.3 mg/m2, R 25 mg (55 pts. (93 %) and 10 mg (4 pts. (7 %)] and Dexamethasone 40 mg/week (59 pts., 100 %). RVd was used in modified schedule in 39 % of pts. in cohort 2 (e.g. RVd weekly, reduced DEX or V dosage). Dose delays were observed in 57 % of pts.In detail cohort 1 (TE) shows a best overall response (ORR)(≥PR) of 93% with 25% PR, 27% VGPR and 41% CR. In comparison in cohort 2 (TNE) the best ORR was 89% with 32% PR, 39% VGPR and 18 % CR.Summary/Conclusion:RVd is a safe and well tolerated therapy over 4 cycles pre‐ASCT and at least 6 cycles in TNE patients. Despite a high dose intensity there were no therapy impeding toxicities. Deep responses (≥VGPR) were frequently achieved (68% in C1, 57% in C2). In TNE pts. dose modifications and delays were frequently observed. Only a minor fraction of patients discontinued therapy [17/141 pts. (12 %)] due to various reasons. In this RW setting RVd proved to be an effective 1st line treatment in both TE and TNE MM patients. Detailed data including relevant side effects and a comparison to clinical trials will be presented.

Quality assessment concept of the World Data Center for Climate and its application to CMIP5 data

Abstract. The preservation of data in a high state of quality which is suitable for interdisciplinary use is one of the most pressing and challenging current issues in long-term archiving. For high volume data such as climate model data, the data and data replica are no longer stored centrally but distributed over several local data repositories, e.g. the data of the Climate Model Intercomparison Project Phase 5 (CMIP5). The most important part of the data is to be archived, assigned a DOI, and published according to the World Data Center for Climate's (WDCC) application of the DataCite regulations. The integrated part of WDCC's data publication process, the data quality assessment, was adapted to the requirements of a federated data infrastructure. A concept of a distributed and federated quality assessment procedure was developed, in which the workload and responsibility for quality control is shared between the three primary CMIP5 data centers: Program for Climate Model Diagnosis and Intercomparison (PCMDI), British Atmospheric Data Centre (BADC), and WDCC. This distributed quality control concept, its pilot implementation for CMIP5, and first experiences are presented. The distributed quality control approach is capable of identifying data inconsistencies and to make quality results immediately available for data creators, data users and data infrastructure managers. Continuous publication of new data versions and slow data replication prevents the quality control from check completion. This together with ongoing developments of the data and metadata infrastructure requires adaptations in code and concept of the distributed quality control approach.

Transfer learning for cross-company software defect prediction

Context Software defect prediction studies usually built models using within-company data, but very few focused on the prediction models trained with cross-company data. It is difficult to employ these models which are built on the within-company data in practice, because of the lack of these local data repositories. Recently, transfer learning has attracted more and more attention for building classifier in target domain using the data from related source domain. It is very useful in cases when distributions of training and test instances differ, but is it appropriate for cross-company software defect prediction? Objective In this paper, we consider the cross-company defect prediction scenario where source and target data are drawn from different companies. In order to harness cross company data, we try to exploit the transfer learning method to build faster and highly effective prediction model. Method Unlike the prior works selecting training data which are similar from the test data, we proposed a novel algorithm called Transfer Naive Bayes (TNB), by using the information of all the proper features in training data. Our solution estimates the distribution of the test data, and transfers cross-company data information into the weights of the training data. On these weighted data, the defect prediction model is built. Results This article presents a theoretical analysis for the comparative methods, and shows the experiment results on the data sets from different organizations. It indicates that TNB is more accurate in terms of AUC (The area under the receiver operating characteristic curve), within less runtime than the state of the art methods. Conclusion It is concluded that when there are too few local training data to train good classifiers, the useful knowledge from different-distribution training data on feature level may help. We are optimistic that our transfer learning method can guide optimal resource allocation strategies, which may reduce software testing cost and increase effectiveness of software testing process.

The Proteios Software Environment: An Extensible Multiuser Platform for Management and Analysis of Proteomics Data

Proteome analysis involves many steps that generate large quantities of data in different formats. This creates a need for automatic data merging and extraction of important features from data. Furthermore, metadata need to be collected and reported to enable critical evaluation of results. Many data analysis tools are developed locally in research laboratories and are nontrivial to adapt for other laboratories, preventing optimal exploitation of generated data. The proteomics field would benefit from user-friendly analysis and data management platforms in which method developers can make their analysis tools available for the community. Here, we describe the Proteios Software Environment (ProSE) that is built around a Web-based local data repository for proteomics experiments. The application features sample tracking, project sharing between multiple users, and automated data merging and analysis. ProSE has built-in support for several quantitative proteomics workflows, and integrates searching in several search engines, automated combination of the search results with predetermined false discovery rates, annotation of proteins and submission of results to public repositories. ProSE also provides a programming interface to enable local extensions, as well as database access using Web services. ProSE provides an analysis platform for proteomics research and is targeted for multiuser projects with needs to share data, sample tracking, and analysis result. ProSE is open source software available at http://www.proteios.org .

Internet based multicenter study for thoracolumbar injuries: a new concept and preliminary results

This article reports about the internet based, second multicenter study (MCS II) of the spine study group (AG WS) of the German trauma association (DGU). It represents a continuation of the first study conducted between the years 1994 and 1996 (MCS I). For the purpose of one common, centralised data capture methodology, a newly developed internet-based data collection system ( http://www.memdoc.org ) of the Institute for Evaluative Research in Orthopaedic Surgery of the University of Bern was used. The aim of this first publication on the MCS II was to describe in detail the new method of data collection and the structure of the developed data base system, via internet. The goal of the study was the assessment of the current state of treatment for fresh traumatic injuries of the thoracolumbar spine in the German speaking part of Europe. For that reason, we intended to collect large number of cases and representative, valid information about the radiographic, clinical and subjective treatment outcomes. Thanks to the new study design of MCS II, not only the common surgical treatment concepts, but also the new and constantly broadening spectrum of spine surgery, i.e. vertebro-/kyphoplasty, computer assisted surgery and navigation, minimal-invasive, and endoscopic techniques, documented and evaluated. We present a first statistical overview and preliminary analysis of 18 centers from Germany and Austria that participated in MCS II. A real time data capture at source was made possible by the constant availability of the data collection system via internet access. Following the principle of an application service provider, software, questionnaires and validation routines are located on a central server, which is accessed from the periphery (hospitals) by means of standard Internet browsers. By that, costly and time consuming software installation and maintenance of local data repositories are avoided and, more importantly, cumbersome migration of data into one integrated database becomes obsolete. Finally, this set-up also replaces traditional systems wherein paper questionnaires were mailed to the central study office and entered by hand whereby incomplete or incorrect forms always represent a resource consuming problem and source of error. With the new study concept and the expanded inclusion criteria of MCS II 1, 251 case histories with admission and surgical data were collected. This remarkable number of interventions documented during 24 months represents an increase of 183% compared to the previously conducted MCS I. The concept and technical feasibility of the MEMdoc data collection system was proven, as the participants of the MCS II succeeded in collecting data ever published on the largest series of patients with spinal injuries treated within a 2 year period.

Semantic data integration in hierarchical domains

A major challenge in building the Semantic Web is resolving differences among heterogeneous databases. This article describes one approach for handling semantic data integration problems in hierarchical domains. It also describes a declarative approach for specifying pairwise mappings between a centrally maintained ontology and each local data repository maintained by an autonomous agency. In this context, it outlines a method for specifying the mappings' semantics and encoding them to resolve heterogeneities. It focuses on XML-based applications in which entities in the centrally maintained ontology are hierarchically related to those in the local data repositories.