Search

Bias could be defined as the tendency to be in favor or against a person or a group, thus promoting unfairness. In computer science, bias is called algorithmic or artificial intelligence (i.e., AI) and can be described as the tendency to showcase recurrent errors in a computer system, which result in “unfair” outcomes. Bias in the “outside world” and algorithmic bias are interconnected since many types of algorithmic bias originate from external factors. The enormous variety of different types of AI biases that have been identified in diverse domains highlights the need for classifying the said types of AI bias and providing a detailed overview of ways to identify and mitigate them. The different types of algorithmic bias that exist could be divided into categories based on the origin of the bias, since bias can occur during the different stages of the Machine Learning (i.e., ML) lifecycle. This manuscript is a literature study that provides a detailed survey regarding the different categories of bias and the corresponding approaches that have been proposed to identify and mitigate them. This study not only provides ready-to-use algorithms for identifying and mitigating bias, but also enhances the empirical knowledge of ML engineers to identify bias based on the similarity that their use cases have to other approaches that are presented in this manuscript. Based on the findings of this study, it is observed that some types of AI bias are better covered in the literature, both in terms of identification and mitigation, whilst others need to be studied more. The overall contribution of this research work is to provide a useful guideline for the identification and mitigation of bias that can be utilized by ML engineers and everyone who is interested in developing, evaluating and/or utilizing ML models.

Abstract Objective Embitterment is a persistent emotion that is known to everybody in reaction to injustice and being let down, associated with feelings of helplessness and hopelessness. People with psychiatric disorders can develop bitterness, which is to be understood as a form of reactive embitterment to the illness. The aim of this explorative study was to investigate the occurrence of embitterment in obsessive–compulsive patients compared to healthy volunteers and in the context of their metacognitions and other biographical and clinical characteristics. Method Following a semi-structured diagnostic interview, a number of measures were administered to 31 patients with obsessive–compulsive disorder (OCD) [ICD-10 F42.X: mean age 35.2 (SD = 10.7) years] and 31 healthy volunteers [mean age 39.1 (SD = 15.0) years]. These measures included the Post-Traumatic Embitterment Disorder questionaire (PTEDq) for measuring embitterment, the Yale-Brown Obsessive–Compulsive Scale, the Metacognition Questionnaire and other psychometric questionnaires such as the Beck Depression Inventory and the State-Trait Anxiety Inventory. Results Patients with OCD scored more than three times higher (mean = 2.0, SD = 1.1) than the healthy participants in the PTEDq (mean = 0.6, SD = 0.8; p

The healthcare sector has been moving toward Electronic Health Record (EHR) systems that produce enormous amounts of healthcare data due to the increased emphasis on getting the appropriate information to the right person, wherever they are, at any time. This highlights the need for a holistic approach to ingest, exploit, and manage these huge amounts of data for achieving better health management and promotion in general. This manuscript proposes such an approach, providing a mechanism allowing all health ecosystem entities to obtain actionable knowledge from heterogeneous data in a multimodal way. The mechanism includes diverse techniques for automatically ingesting healthcare-related information from heterogeneous sources that produce batch/streaming data, managing, fusing, and aggregating this data into new data structures (i.e., Holistic Health Records (HHRs)). The latter enable the aggregation of data coming from different sources, such as Internet of Medical Things (IoMT) devices, online/offline platforms, while to effectively construct the HHRs, the mechanism develops various data management techniques covering the overall data path, from data acquisition and cleaning to data integration, modelling, and interpretation. The mechanism has been evaluated upon different healthcare scenarios, ranging from hospital-retrieved data to patient platforms, combined with data obtained from IoMT devices, having produced useful insights towards its successful and wide adaptation in this domain. In order to implement a paradigm shift from heterogeneous and independent data sources, limited data exploitation, and health records, the mechanism has combined multidisciplinary technologies toward HHRs. Doi: 10.28991/ESJ-2023-07-02-03 Full Text: PDF

Data Cleaning is a subfield of Data Mining that is thriving in the recent years. Ensuring the reliability of data, either when generated or received, is of vital importance to provide the best services possible to users. Accomplishing the aforementioned task is easier said than done, since data are complex, generated at an extremely high rate and are of enormous size. A variety of techniques and methods that are part of other subfields from the domain of the Computer Science have been invoked to assist in making Data Cleaning the most efficient and effective possible. Those subfields include, among others, Natural Language Processing (NLP), which in essence refers to the interaction among computers and human language, seeking to find a way to program computers to be able to process and analyze huge volumes of human language data. NLP is a concept that exists for a long time, but, as time goes by, it is proposed that it can be applied to a variety of concepts that are not solely NLP-related. In this paper, a rule-based data cleaning mechanism is proposed, which utilizes NLP to ensure data reliability. Making use of NLP enabled the mechanism not only to be extremely effective but also to be a lot more efficient compared to other corresponding mechanisms that do not utilize NLP. The mechanism was evaluated upon diverse healthcare datasets, not however being limited to the healthcare domain, but supporting a generalized data cleaning concept.

Product recommendation is considered a well-known technique for bringing customers and products together. With applications in music, electronic shops, or almost any platform the user daily deals with, the recommendation system’s sole scope is to help customers and attract new ones to discover new products. Through product recommendation, transaction costs can also be decreased, improving overall decision-making and quality. To perform recommendations, a recommendation system must utilize customer feedback, such as habits, interests, prior transactions as well as information used in customer profiling, and finally deliver suggestions. Hence, data is the key factor in choosing the appropriate recommendation method and drawing specific suggestions. This research investigates the data challenges of recommendation systems, specifying collaborative-based, content-based, and hybrid-based recommendations. In this context, collaborative filtering is being explored, with the Surprise library and LightFM embeddings being analysed and compared on top of foodservice transactional data. The involved algorithms’ metrics are being identified and parameterized, while hyperparameters are being tuned properly on top of this transactional data, concluding that LightFM provides more efficient recommendation results following the evaluation’s precision and recall outcomes. Nevertheless, even though the Surprise library outperforms, it should be used when constructing user-friendly models, requiring low code and low technicalities. Doi: 10.28991/ESJ-2023-07-01-01 Full Text: PDF

Concerns related to the environmental harm caused by the production, use, and testing of hardware as well as software, are widespread throughout the world. Additionally, there is a problem that the associated subtasks use more power than they ought to. Into this direction, there is a constant drive towards Green Computing since many businesses and organizations are extremely concerned about their carbon footprint and its associated environmental negative effects. The latter is a novel idea that has gained a lot of attention recently and is promoted as a fresh approach towards making computing more environmentally friendly by reducing the impact of computers on the environment. Given that manufacturers, businesses, and individuals have the opportunity to use technology to help the environment through programs and systems aimed at reducing power consumption and saving energy, a practical guide to Green Computing is proposed in this manuscript. This guide offers an overview of Green Computing, a description of its development, as well as a list of its services, depicting the obstacles to its goal. A list of related Green Computing efforts aimed at both large enterprises and individual users, is presented, whereas specific manufacturers and large businesses already performing Green Computing actions are discussed. The aim of this research is to give up-to-date practical guidelines towards a greener and more sustainable environment with reduced carbon emissions that are caused by energy-hungry computing actions and environmentally unaware decisions.

Objectives: Current research aims to address the challenges of exchanging healthcare information, since when this information has to be shared, this happens by specifically designed medical applications or even by the patients themselves. Among the problems that the Health Information Exchange (HIE) initiative is facing are that (i) third party health data cannot be accessed without internet, (ii) there exist crucial delays in accessing citizens’ data, (iii) the direct HIE can only happen among Healthcare Institutions. Methods: Towards the solution of these issues, a Device-to-Device (D2D) protocol has been specified, running on top of the Bluetooth protocol for efficient data exchange. This research is focused on this D2D protocol, by comparing the different Bluetooth profiles that can be used for transmitting this data, based on specific metrics considering the probabilities of transferring erroneous data. Findings: An evaluation of three Bluetooth profiles takes place, concluding that two of the three profiles must be used to respect the D2D protocol nature and be fully supported by the main market vendors’ operating systems. Novelty:Based on this evaluation, the specified D2D protocol has been built on top of state-of-the-art short-range distance communication technologies, fully supporting the healthcare ecosystem towards the HIE paradigm. Doi: 10.28991/esj-2021-01276 Full Text: PDF

Electronic Health Record (EHR) data is being required to be shaped into a standardized and interoperable format for easy and secure accessing from anywhere, towards the vision of a borderless Health Information Exchange (HIE). Towards this vision, EHR data is being securely stored in different cloud repositories, making it however difficult and time-consuming to access it in emergency cases from healthcare practitioners, or even from the owners. In order to access the EHR data, it is needed specific credentials and real-time approved consent, which cannot be feasible in cases where this data must be urgently accessed and furtherly processed. In this paper, a health record indexing methodology is being proposed (i.e. HR index) that facilitates the real-time accessing of non-privileged users, to the cloud stored EHR data of citizens who are facing emergency situations. In this context, the research that has been conducted prior to specifying and designing the proposed HR index is being provided, including a detailed study of data and health data indexing techniques and methodologies. The overall vision, scope and usage of the HR index is provided in the context of an Emergency scenario, concluding to the design and evaluation of the HR index, under specific circumstances and requirements.

We present PolicyCLOUD: a prototype for an extensible serverless cloud-based system that supports evidence-based elaboration and analysis of policies. PolicyCLOUD allows flexible exploitation and management of policy-relevant dataflows, by enabling the practitioner to register datasets and specify a sequence of transformations and/or information extraction through registered ingest functions. Once a possibly transformed dataset has been ingested, additional insights can be retrieved by further applying registered analytic functions to it. PolicyCLOUD was built as an extensible framework toward the creation of an analytic ecosystem. As of now, we have developed several essential ingest and analytic functions that are built-in within the framework. They include data cleaning, enhanced interoperability, and sentiment analysis generic functions; in addition, a trend analysis function is being created as a new built-in function. PolicyCLOUD has also the ability to tap on the analytic capabilities of external tools; we demonstrate this with a social dynamics tool implemented in conjunction with PolicyCLOUD, and describe how this stand-alone tool can be integrated with the PolicyCLOUD platform to enrich it with policy modeling, design and simulation capabilities. Furthermore, PolicyCLOUD is supported by a tailor-made legal and ethical framework derived from privacy/data protection best practices and existing standards at the EU level, which regulates the usage and dissemination of datasets and analytic functions throughout its policy-relevant dataflows. The article describes and evaluates the application of PolicyCLOUD to four families of pilots that cover a wide range of policy scenarios.

Abstract Background The experience of time, or the temporal order of external and internal events, is essential for humans. In psychiatric disorders such as depression and schizophrenia, impairment of time processing has been discussed for a long time. Aims In this explorative pilot study, therefore, the subjective time feeling as well as objective time perception were determined in patients with depression and schizophrenia, along with possible neurobiological correlates. Methods Depressed (n = 34; 32.4 ± 9.8 years; 21 men) and schizophrenic patients (n = 31; 35.1 ± 10.7 years; 22 men) and healthy subjects (n = 33; 32.8 ± 14.3 years; 16 men) were tested using time feeling questionnaires, time perception tasks and critical flicker-fusion frequency (CFF) and loudness dependence of auditory evoked potentials (LDAEP) to determine serotonergic neurotransmission. Results There were significant differences between the three groups regarding time feeling and also in time perception tasks (estimation of given time duration) and CFF (the “DOWN” condition). Regarding the LDAEP, patients with schizophrenia showed a significant negative correlation to time experience in terms of a pathologically increased serotonergic neurotransmission with disturbed time feeling. Conclusions Impairment of time experience seems to play an important role in depression and schizophrenia, both subjectively and objectively, and novel neurobiological correlates have been uncovered. It is suggested, therefore, that alteration of experience of time should be increasingly included in the current psychopathological findings.

Big Data is a phenomenon that affects today’s world, with new data being generated every second. Today’s enterprises face major challenges from the increasingly diverse data, as well as from indexing, searching, and analyzing such enormous amounts of data. In this context, several frameworks and libraries for processing and analyzing Big Data exist. Among those frameworks Hadoop MapReduce, Mahout, Spark, and MLlib appear to be the most popular, although it is unclear which of them best suits and performs in various data processing and analysis scenarios. This paper proposes EverAnalyzer, a self-adjustable Big Data management platform built to fill this gap by exploiting all of these frameworks. The platform is able to collect data both in a streaming and in a batch manner, utilizing the metadata obtained from its users’ processing and analytical processes applied to the collected data. Based on this metadata, the platform recommends the optimum framework for the data processing/analytical activities that the users aim to execute. To verify the platform’s efficiency, numerous experiments were carried out using 30 diverse datasets related to various diseases. The results revealed that EverAnalyzer correctly suggested the optimum framework in 80% of the cases, indicating that the platform made the best selections in the majority of the experiments.

Extracting useful knowledge from proper data analysis is a very challenging task for efficient and timely decision-making. To achieve this, there exist a plethora of machine learning (ML) algorithms, while, especially in healthcare, this complexity increases due to the domain’s requirements for analytics-based risk predictions. This manuscript proposes a data analysis mechanism experimented in diverse healthcare scenarios, towards constructing a catalogue of the most efficient ML algorithms to be used depending on the healthcare scenario’s requirements and datasets, for efficiently predicting the onset of a disease. To this context, seven (7) different ML algorithms (Naïve Bayes, K-Nearest Neighbors, Decision Tree, Logistic Regression, Random Forest, Neural Networks, Stochastic Gradient Descent) have been executed on top of diverse healthcare scenarios (stroke, COVID-19, diabetes, breast cancer, kidney disease, heart failure). Based on a variety of performance metrics (accuracy, recall, precision, F1-score, specificity, confusion matrix), it has been identified that a sub-set of ML algorithms are more efficient for timely predictions under specific healthcare scenarios, and that is why the envisioned ML catalogue prioritizes the ML algorithms to be used, depending on the scenarios’ nature and needed metrics. Further evaluation must be performed considering additional scenarios, involving state-of-the-art techniques (e.g., cloud deployment, federated ML) for improving the mechanism’s efficiency.

Background: Mothers with postpartum depression (PPD) show impaired affects and behaviour patterns in the mother-child interaction, which affects an infant's emotional and cognitive development and the maternal course of disease. However, impairment of the mother-child relationship does not occur in every case of PPD. Aim: The aim of this exploratory-descriptive video-based study was to investigate the possible associations between mother-child interactions and aspects of maternal biography and clinical history, with a focus on pre-existing mental disorder. Methods: Sixty-two mother-child dyads (31 mothers with PPD and pre-existing mental disorders and 31 mothers with PPD but no further mental disorder) hospitalized at the mother and baby unit (MBU) of the LWL-Hospital Herten were included in this study. The Marcé Clinical Checklist and the “Mannheimer Beurteilungsskala zur Erfassung der Mutter-Kind-Interaktion im Säuglingsalter” (MBS-MKI-S) were used to explore sociodemographic and clinical parameters, and video-based interaction behaviour was examined. Results: Mother-infant interaction behaviour showed a significant group difference on the MBS-MKI-S-Vm subscale (variability in maternal behaviour) before psychiatric treatment (exact Mann-Whitney U test: U = 555, p = 0.023), with higher scores in mothers with a pre-existing mental disorder. Furthermore, significant differences were shown on the MBS-MKI-S-RSm (maternal reactivity/sensitivity) (U = 259, p = 0.019) and MBS-MKI-S-Rc (child's reactivity) subscales at discharge (U = 251, p = 0.021). Among mothers with a pre-existing diagnosis, the MBS-MKI-S-Tm (maternal tenderness) and MBS-MKI-S-Rc (child's reactivity) subscales were significantly correlated after treatment. Conclusions: Mothers with PPD and a pre-existing mental disorder displayed significantly more behavioural variability than mothers with only PPD. Maternal behaviour seems to influence the child's responsive behaviour; thus, mothers and their children can benefit from inpatient treatment at an MBU. Further investigations with larger samples should be conducted.

Nowadays, the use of Internet of Things (IoT) in various different fields has made significant progress, especially in the field of healthcare, where a myriad of heterogeneous medical data sources are in use. This fact has reinforced the vision of developing new communication technologies and finding new ways to synchronize and successfully manage all these data sources. However, this vision is accompanied by several related challenges. One of these challenges refers to the fact that since all the existing IoT medical data sources are usually characterized by a high degree of heterogeneity, they are expected to be recognized as reliable at different stages, thus providing data of different levels of reliability. To effectively tackle this challenge, the present paper proposes a mechanism for capturing the reliability levels of different IoT medical data sources, so as to automatically decide whether these will be considered as reliable or not, and thus their data will be kept for further analysis. In this context, in this mechanism three (3) discrete stages are implemented, facilitating both the data reliability and the availability estimation of these data sources, making finally feasible the manipulation of these sources and the estimation of their overall reliability levels. The prototype associated with this paper provides an example of this mechanism, demonstrating in detail each discrete stage.

The Healthcare 4.0 era is surrounded by challenges varying from the Internet of Medical Things (IoMT) devices’ data collection, integration and interpretation. Several techniques have been developed that however do not propose solutions that can be applied to different scenarios or domains. When dealing with healthcare data, based on the severity and the application of their results, they should be provided almost in real-time, without any errors, inconsistencies or misunderstandings. Henceforth, in this manuscript a platform is proposed for efficiently managing healthcare data, by taking advantage of the latest techniques in Data Acquisition, 5G Network Slicing and Data Interoperability. In this platform, IoMT devices’ data and network specifications can be acquired and segmented in different 5G network slices according to the severity and the computation requirements of different medical scenarios. In sequel, transformations are performed on the data of each network slice to address data heterogeneity issues, and provide the data of the same network slices into HL7 FHIR-compliant format, for further analysis.

Over the past few years, increasing attention has been given to the health sector and the integration of new technologies into it. Cloud computing and storage clouds have become essentially state of the art solutions for other major areas and have started to rapidly make their presence powerful in the health sector as well. More and more companies are working toward a future that will allow healthcare professionals to engage more with such infrastructures, enabling them a vast number of possibilities. While this is a very important step, less attention has been given to the citizens. For this reason, in this paper, a citizen-centered storage cloud solution is proposed that will allow citizens to hold their health data in their own hands while also enabling the exchange of these data with healthcare professionals during emergency situations. Not only that, in order to reduce the health data transmission delay, a novel context-aware prefetch engine enriched with deep learning capabilities is proposed. The proposed prefetch scheme, along with the proposed storage cloud, is put under a two-fold evaluation in several deployment and usage scenarios in order to examine its performance with respect to the data transmission times, while also evaluating its outcomes compared to other state of the art solutions. The results show that the proposed solution shows significant improvement of the download speed when compared with the storage cloud, especially when large data are exchanged. In addition, the results of the proposed scheme evaluation depict that the proposed scheme improves the overall predictions, considering the coefficient of determination (R2 > 0.94) and the mean of errors (RMSE < 1), while also reducing the training data by 12%.

Background: Death and the anxiety of it becomes more apparent when confronted with a chronic disease. Even though multiple sclerosis (MS) is a treatable condition today, it is still accompanied by a multitude of impairments, which in turn may intensify of death anxiety. Objective: The aim of this study is to explore the relationship between depression, anxiety and death anxiety in individuals with MS. Methods: Fifty-six MS patients were recruited at the Department of Neurology of the University Clinic in Basel. Death anxiety was assessed using the Bochumer Questionnaire on attitude to death and death anxiety 2.0 (BOFRETTA 2.0). Results: Scores of death anxiety towards it in MS patients were low. Only disability (EDSS) was moderately correlated with death anxiety. Depression in MS was significantly correlated with fatigue and disability, but not with the BOFRETTA 2.0. Conclusion: Scores of death anxiety and the attitude towards death are low in this MS cohort. It was shown that both psychopathological and neurological deficits impact the subject of death with respect to multiple sclerosis.

Background: Given the lethal severity of amyotrophic lateral sclerosis (ALS), the aim of this study was to illuminate the coherence of depression and death anxiety in both ALS patients and caregivers and in how far patients and caregivers are influenced by the mindset of their respective counterpart.Methods: 30 couples of patients (mean age 60.57; 13 women, 17 men) and primary caregivers (mean age 57.33; 16 women, 14 men) were included into the study. Death anxiety was assessed using the newly developed BOFRETTA scale, depression via Beck Depression Inventory, anxiety by means of State Trait Anxiety Inventory and caregivers' exertion using the Caregiver Strain Index. Patients' impairment was assessed with the ALS functional rating scale.Results: We found that while death anxiety was related to depression in both patients and caregivers, death anxiety was related to anxiety only in patients. Caregiver strain correlated with both caregiver‘s depression and anxiety. Moreover, patients' and caregivers’ depression, anxiety and death anxiety correlated to the ones of their counterpart.Conclusion: These results suggest that despite little depressive symptoms in ALS patients the fatal prognosis of the disease takes into account, depression and death anxiety influence each other and might be addressed together in pharmacological and especially psychotherapeutic interventions to the benefit of the patient. Medical professionals should not forget to offer sufficient support to caregivers tending patients affected by depression and death anxiety as they are likely to mirror their patient's feelings.

It is an undeniable fact that Internet of Things (IoT) technologies have become a milestone advancement in the digital healthcare domain, since the number of IoT medical devices is grown exponentially, and it is now anticipated that by 2020 there will be over 161 million of them connected worldwide. Therefore, in an era of continuous growth, IoT healthcare faces various challenges, such as the collection, the quality estimation, as well as the interpretation and the harmonization of the data that derive from the existing huge amounts of heterogeneous IoT medical devices. Even though various approaches have been developed so far for solving each one of these challenges, none of these proposes a holistic approach for successfully achieving data interoperability between high-quality data that derive from heterogeneous devices. For that reason, in this manuscript a mechanism is produced for effectively addressing the intersection of these challenges. Through this mechanism, initially, the collection of the different devices’ datasets occurs, followed by the cleaning of them. In sequel, the produced cleaning results are used in order to capture the levels of the overall data quality of each dataset, in combination with the measurements of the availability of each device that produced each dataset, and the reliability of it. Consequently, only the high-quality data is kept and translated into a common format, being able to be used for further utilization. The proposed mechanism is evaluated through a specific scenario, producing reliable results, achieving data interoperability of 100% accuracy, and data quality of more than 90% accuracy.