Progression to myocardial infarction short-term death based on interval sequential pattern mining.

BACKGROUND: Myocardial infarction (MI) is one of the significant cardiovascular diseases (CVDs). According to Taiwanese health record analysis, the hazard rate reaches a peak in the initial year after diagnosis of MI, drops to a relatively low value, and maintains stable for the following years. Therefore, identifying suspicious comorbidity patterns of short-term death before the diagnosis may help achieve prolonged survival for MI patients. METHODS: Interval sequential pattern mining was applied with odds ratio to the hospitalization records from the Taiwan National Health Insurance Research Database to evaluate the disease progression and identify potential subjects at the earliest possible stage. RESULTS: Our analysis resulted in five disease pathways, including "diabetes mellitus," "other disorders of the urethra and urinary tract," "essential hypertension," "hypertensive heart disease," and "other forms of chronic ischemic heart disease" that led to short-term death after MI diagnosis, and these pathways covered half of the cohort. CONCLUSION: We explored the possibility of establishing trajectory patterns to identify the high-risk population of early mortality after MI.

Framework to Detect Schizophrenia in Brain MRI Slices with Mayfly Algorithm-Selected Deep and Handcrafted Features.

Brain abnormality causes severe human problems, and thorough screening is necessary to identify the disease. In clinics, bio-image-supported brain abnormality screening is employed mainly because of its investigative accuracy compared with bio-signal (EEG)-based practice. This research aims to develop a reliable disease screening framework for the automatic identification of schizophrenia (SCZ) conditions from brain MRI slices. This scheme consists following phases: (i) MRI slices collection and pre-processing, (ii) implementation of VGG16 to extract deep features (DF), (iii) collection of handcrafted features (HF), (iv) mayfly algorithm-supported optimal feature selection, (v) serial feature concatenation, and (vi) binary classifier execution and validation. The performance of the proposed scheme was independently tested with DF, HF, and concatenated features (DF+HF), and the achieved outcome of this study verifies that the schizophrenia screening accuracy with DF+HF is superior compared with other methods. During this work, 40 patients' brain MRI images (20 controlled and 20 SCZ class) were considered for the investigation, and the following accuracies were achieved: DF provided >91%, HF obtained >85%, and DF+HF achieved >95%. Therefore, this framework is clinically significant, and in the future, it can be used to inspect actual patients' brain MRI slices.

Privacy protection framework for face recognition in edge-based Internet of Things.

Edge computing (EC) gets the Internet of Things (IoT)-based face recognition systems out of trouble caused by limited storage and computing resources of local or mobile terminals. However, data privacy leak remains a concerning problem. Previous studies only focused on some stages of face data processing, while this study focuses on the privacy protection of face data throughout its entire life cycle. Therefore, we propose a general privacy protection framework for edge-based face recognition (EFR) systems. To protect the privacy of face images and training models transmitted between edges and the remote cloud, we design a local differential privacy (LDP) algorithm based on the proportion difference of feature information. In addition, we also introduced identity authentication and hash technology to ensure the legitimacy of the terminal device and the integrity of the face image in the data acquisition phase. Theoretical analysis proves the rationality and feasibility of the scheme. Compared with the non-privacy protection situation and the equal privacy budget allocation method, our method achieves the best balance between availability and privacy protection in the numerical experiment.

Monitoring moving queries inside a safe region.

With mobile moving range queries, there is a need to recalculate the relevant surrounding objects of interest whenever the query moves. Therefore, monitoring the moving query is very costly. The safe region is one method that has been proposed to minimise the communication and computation cost of continuously monitoring a moving range query. Inside the safe region the set of objects of interest to the query do not change; thus there is no need to update the query while it is inside its safe region. However, when the query leaves its safe region the mobile device has to reevaluate the query, necessitating communication with the server. Knowing when and where the mobile device will leave a safe region is widely known as a difficult problem. To solve this problem, we propose a novel method to monitor the position of the query over time using a linear function based on the direction of the query obtained by periodic monitoring of its position. Periodic monitoring ensures that the query is aware of its location all the time. This method reduces the costs associated with communications in client-server architecture. Computational results show that our method is successful in handling moving query patterns.

Analysing the impact of comorbid conditions and media coverage on online symptom search data: a novel AI-based approach for COVID-19 tracking.

BACKGROUND: Web search data have proven to bea valuable early indicator of COVID-19 outbreaks. However, the influence of co-morbid conditions with similar symptoms and the effect of media coverage on symptom-related searches are often overlooked, leading to potential inaccuracies in COVID-19 simulations. METHOD: This study introduces a machine learning-based approach to estimate the magnitude of the impact of media coverage and comorbid conditions with similar symptoms on online symptom searches, based on two scenarios with quantile levels 10-90 and 25-75. An incremental batch learning RNN-LSTM model was then developed for the COVID-19 simulation in Australia and New Zealand, allowing the model to dynamically simulate different infection rates and transmissibility of SARS-CoV-2 variants. RESULT: The COVID-19 infected person-directed symptom searches were found to account for only a small proportion of the total search volume (on average 33.68% in Australia vs. 36.89% in New Zealand) compared to searches influenced by media coverage and comorbid conditions (on average 44.88% in Australia vs. 50.94% in New Zealand). The proposed method, which incorporates estimated symptom component ratios into the RNN-LSTM embedding model, significantly improved COVID-19 simulation performance. CONCLUSION: Media coverage and comorbid conditions with similar symptoms dominate the total number of online symptom searches, suggesting that direct use of online symptom search data in COVID-19 simulations may overestimate COVID-19 infections. Our approach provides new insights into the accurate estimation of COVID-19 infections using online symptom searches, thereby assisting governments in developing complementary methods for public health surveillance.

Making the most of what we have: What does the future hold for Emergency Department data?

Over 10 million ED visits occur each year across Australia and Aotearoa New Zealand. Outside basic administrative data focused on time-based targets, there is minimal information about clinical performance, quality of care, patient outcomes, or equity in emergency care. The lack of a timely, accurate or clinically useful data collection represents a missed opportunity to improve the care we deliver each day. The present paper outlines a proposal for a National Acute Care Secure Health Data Environment, including design, possible applications, and the steps taken to date by the Australasian College for Emergency Medicine ED Epidemiology Network in collaboration with the College of Emergency Nursing Australasia. Optimal use of the existing information collected routinely during clinical care of emergency patients has the potential to enable data-driven quality improvement and research, leading to better care and better outcomes for millions of patients and families each year.

Improved cloud storage auditing scheme with deduplication.

Cloud storage has become a crucial service for many users who deal with big data. The auditing scheme for cloud storage is a mechanism that checks the integrity of outsourced data. Cloud storage deduplication is a technique that helps cloud service providers save on storage costs by storing only one copy of a file when multiple users outsource the same file to cloud servers. However, combining storage auditing and deduplication techniques can be challenging. To address this challenge, in 2019 Hou et al. proposed a cloud storage auditing scheme with deduplication that supports different security levels of data popularity. This proposal is interesting and has practical applications. However, in this paper, we show that their proposal has a flaw: the cloud or other adversaries can easily forge the data block's authenticators, which means the cloud can delete all the outsourced encrypted data blocks but still provide correct storage proof for the third-party auditor. Based on Hou et al.'s scheme, we propose an improved cloud storage auditing scheme with deduplication and analyze its security. The results show that the proposed scheme is more secure.

COVID-19 Prevention Strategies for Victoria Students within Educational Facilities: An AI-Based Modelling Study.

Educational institutions play a significant role in the community spread of SARS-CoV-2 in Victoria. Despite a series of social restrictions and preventive measures in educational institutions implemented by the Victorian Government, confirmed cases among people under 20 years of age accounted for more than a quarter of the total infections in the state. In this study, we investigated the risk factors associated with COVID-19 infection within Victoria educational institutions using an incremental deep learning recurrent neural network-gated recurrent unit (RNN-GRU) model. The RNN-GRU model simulation was built based on three risk dimensions: (1) school-related risk factors, (2) student-related community risk factors, and (3) general population risk factors. Our data analysis showed that COVID-19 infection cases among people aged 10-19 years were higher than those aged 0-9 years in the Victorian region in 2020-2022. Within the three dimensions, a significant association was identified between school-initiated contact tracing (0.6110), vaccination policy for students and teachers (0.6100), testing policy (0.6109), and face covering (0.6071) and prevention of COVID-19 infection in educational settings. Furthermore, the study showed that different risk factors have varying degrees of effectiveness in preventing COVID-19 infection for the 0-9 and 10-19 age groups, such as state travel control (0.2743 vs. 0.3390), international travel control (0.2757 vs. 0.3357) and school closure (0.2738 vs. 0.3323), etc. More preventive support is suggested for the younger generation, especially for the 10-19 age group.

Detection of Amyotrophic Lateral Sclerosis (ALS) Comorbidity Trajectories Based on Principal Tree Model Analytics.

The multifaceted nature and swift progression of Amyotrophic Lateral Sclerosis (ALS) pose considerable challenges to our understanding of its evolution and interplay with comorbid conditions. This study seeks to elucidate the temporal dynamics of ALS progression and its interaction with associated diseases. We employed a principal tree-based model to decipher patterns within clinical data derived from a population-based database in Taiwan. The disease progression was portrayed as branched trajectories, each path representing a series of distinct stages. Each stage embodied the cumulative occurrence of co-existing diseases, depicted as nodes on the tree, with edges symbolizing potential transitions between these linked nodes. Our model identified eight distinct ALS patient trajectories, unveiling unique patterns of disease associations at various stages of progression. These patterns may suggest underlying disease mechanisms or risk factors. This research re-conceptualizes ALS progression as a migration through diverse stages, instead of the perspective of a sequence of isolated events. This new approach illuminates patterns of disease association across different progression phases. The insights obtained from this study hold the potential to inform doctors regarding the development of personalized treatment strategies, ultimately enhancing patient prognosis and quality of life.

MCGNet+: an improved motor imagery classification based on cosine similarity.

It has been a challenge for solving the motor imagery classification problem in the brain informatics area. Accuracy and efficiency are the major obstacles for motor imagery analysis in the past decades since the computational capability and algorithmic availability cannot satisfy complex brain signal analysis. In recent years, the rapid development of machine learning (ML) methods has empowered people to tackle the motor imagery classification problem with more efficient methods. Among various ML methods, the Graph neural networks (GNNs) method has shown its efficiency and accuracy in dealing with inter-related complex networks. The use of GNN provides new possibilities for feature extraction from brain structure connection. In this paper, we proposed a new model called MCGNet+, which improves the performance of our previous model MutualGraphNet. In this latest model, the mutual information of the input columns forms the initial adjacency matrix for the cosine similarity calculation between columns to generate a new adjacency matrix in each iteration. The dynamic adjacency matrix combined with the spatial temporal graph convolution network (ST-GCN) has better performance than the unchanged matrix model. The experimental results indicate that MCGNet+ is robust enough to learn the interpretable features and outperforms the current state-of-the-art methods.

Where do low risk women live relative to maternity services across Victoria? Expanding access to public homebirth models across Victoria.

PROBLEM: Currently <1% of Australian women give birth at home. BACKGROUND: In Australia there are very few options for women to access public funded homebirth. AIM: We aimed to use geo-mapping to identify the number of women eligible for homebirth in Victoria, based on the criteria of uncomplicated pregnancies and residing within 15-25kms of suitable maternity services, to plan future maternity care options. METHODS: Retrospective study of births between 2015 and 2017 in Victoria, Australia. All women who were identified as having a low risk pregnancy at the beginning of pregnancy were included. The number of women within 15 and 25km of a suitable Victorian public maternity hospital and catchment boundaries around each hospital were determined. FINDINGS: Between 2015 and 2017, 126,830 low risk women gave birth in Victoria, of whom half live within 25km of seven Victorian hospitals. Currently, 2% of suitable women who live close to the current public homebirth models accessed them. DISCUSSION: We present a method to inform the expansion of maternity service options using Victoria as an example. On the basis of the maximum number of low risk women living close by, we have also identified the Victorian maternity services that would be most suitable for creation of public homebirth or low risk continuity of midwifery models. CONCLUSION: This approach could can be used to plan other maternity care services.

VGG19 Network Assisted Joint Segmentation and Classification of Lung Nodules in CT Images.

Pulmonary nodule is one of the lung diseases and its early diagnosis and treatment are essential to cure the patient. This paper introduces a deep learning framework to support the automated detection of lung nodules in computed tomography (CT) images. The proposed framework employs VGG-SegNet supported nodule mining and pre-trained DL-based classification to support automated lung nodule detection. The classification of lung CT images is implemented using the attained deep features, and then these features are serially concatenated with the handcrafted features, such as the Grey Level Co-Occurrence Matrix (GLCM), Local-Binary-Pattern (LBP) and Pyramid Histogram of Oriented Gradients (PHOG) to enhance the disease detection accuracy. The images used for experiments are collected from the LIDC-IDRI and Lung-PET-CT-Dx datasets. The experimental results attained show that the VGG19 architecture with concatenated deep and handcrafted features can achieve an accuracy of 97.83% with the SVM-RBF classifier.