List Of Features Research Articles

Personalized Privacy Assistant: Identity Construction and Privacy in the Internet of Things.

Over time, the many different ways in which we collect and use data have become more complex as we communicate and interact with an ever-increasing variety of modern technologies. Although people often say they care about their privacy, they do not have a deep understanding of what devices around them are collecting their identity information, what identity information is being collected, and how that collected data will affect them. This research is dedicated to developing a personalized privacy assistant to help users regain control, understand their own identity management, and process and simplify the large amount of information from the Internet of Things (IoT). This research constructs an empirical study to obtain the comprehensive list of identity attributes that are being collected by IoT devices. We build a statistical model to simulate the identity theft and to help calculate the privacy risk score based on the identity attributes collected by IoT devices. We discuss how well each feature of our Personal Privacy Assistant (PPA) works and compare the PPA and related work to a list of fundamental features for privacy protection.

Deep learning based spatio‐temporal hand gesture recognition system in complex environment

AbstractGesture recognition nowadays has grabbed the attention of researchers as they represent human behaviour in multiple practical ways. Amongst a variety of gestures available, hand gestures play an essential role in the field of human‐computer interaction when recognised efficiently in complex and dynamic environments. In this paper, we propose a dynamic hand gesture recognition system to recognise hand gestures appearing in different indoor and outdoor environments. Hand detection and tracking uses a two‐level system resulting in the formation of gesture trajectory in challenging conditions in which existing detection and tracking algorithms could not do so. A set of 45 features is provided as input to the various classification techniques. The redundancy problem has been reduced by selecting a set of optimum features using the analysis of variance method, which ranks the list of features. An incremental feature selection technique calculates recognition accuracy by selecting features according to rankings. This system provides an accuracy of 96.32% when used with machine learning and 97.5% when used with deep learning techniques. Recognition accuracy is calculated for various environments, including an extra hand, multiple persons in the video frame, and outdoor environment. All machine‐learning classifiers are combined using classifier combination to calculate the accuracy according to the majority‐voting rule. Based on the experimental results, it has been observed that deep learning provides better results compared to machine learning.

Adalimumab originator versus adalimumab biosimilars in inflammatory bowel disease in Australia

ABSTRACT Background Biosimilar adalimumabs have improved treatment access, but without any clinical advantage, distributors rely on delivery device design-enhancements, support services, and removal of painful excipients to capture market share. Prescribers, however, are often unaware of these differences. This article compares and contrasts originator versus biosimilar adalimumab agents to identify key differences that might influence adalimumab selection. Research design and methods We reviewed listed adalimumab biosimilars in Australia and compared them to the originator adalimumab. Similarities and differences identified were confirmed with the manufacturers via two rounds of interviews: the first to collate a list of features and benefits of their product, and the second to consolidate and confirm the data. Results The originator adalimumab Humira [by AbbVie, U.S.A] and four adalimumab biosimilars (Amgevita [by Amgen, U.S.A], Hadlima [by Organon, U.S.A], Hyrimoz [by Sandoz, Switzerland], and Idacio [by Fresenius Kabi, Germany]) are included in this review. Key differences identified include product formulation, dosages available, delivery devices, physician support, patient support, and the supply of other biosimilar products by the company. Conclusion Adalimumab biosimilars are different from each other with unique advantages and disadvantages likely to influence prescriber and patients. Therefore, the choice of agent should be individualized to the needs of the patient and the healthcare service.

Building machine-learning models for reducing the severity of bicyclist road traffic injuries

Predicting the severity of injuries caused by traffic accidents is an important undertaking because it may lead to establishing regulations increasing road-user safety. Bicyclists are a particularly susceptible category of road users, which is especially troubling considering the environmental, financial, and health benefits of this mode of transportation. As a result, this study aims to apply machine learning to identify risk variables that may result in serious biker injuries in the case of an accident.Machine-learning models make no assumptions about the connections between variables. Hence, it has been argued that machine-learning approaches produce better outcomes than statistical procedures. This study selects the “best” machine-learning classification system from a vast pool of similar algorithms to predict the severity of bicycling injuries. Machine learning allows the system to learn from experience and improve without being too programmed.We first use a variety of feature selection algorithms to identify a list of features related to the accident and the environment that have the greatest impact on the severity of bicyclist injuries. This feature list is then used as input data to various machine-learning algorithms that predict the class of bicyclist injury severity at one of three levels (fatal, serious, and slight). The “best” machine-learning algorithm is identified on the basis of having the highest levels of accuracy, precision, recall, and F1 score. The current models were developed and trained based on Israeli road-traffic accident data from 2009 to 2019, meaning that new models would need to be developed for other geographical locations. In addition, the models would need to be updated to take account of the changing relationships between motorists, bicyclists, and the environment. Nevertheless, the proposed methodology has universal applicability.

Enhancing multilayer perceptron neural network using archive-based harris hawks optimizer to predict gold prices

The success of the Multi-Layer Perceptron Neural Network (MLP) relies on carefully configuring its weights and biases to promising values. The gradient descent technique is usually used to select the optimal MLP configuration. Such a technique can stuck in local optima and its convergence towards promising regions in the search space is slow. Harris Hawks optimizer (HHO) is a recently proposed metaheuristic technique that mimics the harris’ hawks behavior to chase and catch prey. In this paper, the algorithm search process is enhanced by introducing an external archive that saves the best solutions to be used in the next iterations. Gold price prediction is essential for future investment planning and decision-making in mining projects. The gold price is predicted using a multilayer perceptron (MLP) neural network (NN) that has the optimal control parameters obtained using the improved version of Harris Hawks optimizer (HHO) with an external archive which is called (AHHO-NN). The input features with a complicated relationship can influence gold price prediction models such as inflation rate, precious metals prices, and other features. In this sense, the literature suggests a specific reduced number of features for the MLP-NN model to yield high-quality prediction results. In addition to the prediction model, two feature reduction methods are used to select the proper list of the input features for the MLP-NN model to forecast gold price namely, Pearson’s correlation and a newly proposed categorized correlation. Surprisingly, new features not mentioned in the literature are discovered and some are discarded. The time series dataset used has been extracted from several sources and pre-processed to fit the proposed model. Furthermore, the prediction results have been evaluated using several measures such as prediction accuracy, Mean square error (MSE), Root mean square error (RMSE), Mean absolute error (MAE), and average absolute relative deviation (AARD). The performance of the proposed AHHO-NN model is compared against four swarm intelligence algorithms (i.e., HHO-NN, JAYA-NN, MFO-NN, PSO-NN) as well as four classical machine learning techniques (i.e., linear regression (LR), MLP Regressor (MLP), RANSAC Regressor (RANSAC), and TheilSen Regressor (TR)). Interestingly, the proposed AHHO-NN is able to comparatively produce high-quality prediction results for gold prices with relatively high accuracy. This proves the viability of the proposed AHHO-NN model as well as the proper selection of the primitive input features.

Abstract LB056: Crowdsourcing rare cancer research in the Hack4NF GENIE-NF tumor identification and classification challenge

Abstract A key challenge in rare tumor research is the paucity of genomic data that can be used to understand and devise better therapeutic strategies for rare cancers. Furthermore, the “curse of dimensionality,” in which data has many features, such as genetic variants, but few specimens, makes it difficult or impossible to use conventional machine learning techniques to explore these data. To address these challenges in the context of tumors associated with the rare disease neurofibromatosis, we ran a hackathon to stimulate the development of methods to better understand the biology of tumors related to this disease. The hackathon had three challenges centered around variant effect prediction, drug discovery, and genomics. The genomics track of the hackathon leveraged the AACR Project GENIE database (1) and challenged participants to develop new frameworks that accurately use GENIE data to classify neurofibromatosis-related tumors. They were asked to first identify the neurofibromatosis-related tumors in the dataset. They were then asked to use one or more novel classification methods to classify the tumor samples into different groups based on genetic features. To help them do this, we provided access to version 13 of the GENIE database to the hackathon participants, though they were allowed to integrate other relevant datasets. The expected output was a classification method that differentiates different types of NF1, NF2, and schwannomatosis-related tumors using clinical sequencing data, as well as a list of the most important features in the algorithm for differentiating tumor types. Domain expert judges qualitatively scored each team’s rationale for defining and including “NF-related tumors” in their project, and scored the feature list based on the presence of known important biomarkers and features in NF tumors as well as potentially novel features that the algorithm identified. A technical judge also scored the code repository based on documentation and clarity of code. Two teams from the GENIE subchallenge were awarded prizes - Team Next GeNLP as the best overall GENIE challenge submission, and team “Artificial Intelligence for neurofibromatosis” for best project documentation. Both winning teams used methods based on natural language processing (NLP) techniques to reduce the dimensionality and complexity of the variant data, and to identify new representations of NF-relevant tumors, and then applied downstream analysis methods such as distance calculations and feature prioritization to better understand the genomic profiles of different tumors. While these methods and tools focused on NF-specific tumor types, we anticipate that they could be re-used by others to better explore the biology and interrelatedness of other rare tumors within the GENIE database. (1) The AACR Project GENIE Consortium. AACR Project GENIE: Powering Precision Medicine Through An International Consortium, Cancer Discov. 2017. The authors would like to acknowledge the American Association for Cancer Research and its financial and material support in the development of the AACR Project GENIE registry, as well as members of the consortium for their commitment to data sharing. Interpretations are the responsibility of study authors. Citation Format: Robert J. Allaway, Sasha Scott, Gabriel Altay, Hariprasad Donthi, Karthika R, Muhammad Alaa Alwattar, Lucas Pastur Romay, Ayesha Parsha, Jineta Banerjee, Chelsea Nayan, Julie Bletz, Salvatore La Rosa. Crowdsourcing rare cancer research in the Hack4NF GENIE-NF tumor identification and classification challenge [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 2 (Clinical Trials and Late-Breaking Research); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(8_Suppl):Abstract nr LB056.

The CellPhe toolkit for cell phenotyping using time-lapse imaging and pattern recognition

With phenotypic heterogeneity in whole cell populations widely recognised, the demand for quantitative and temporal analysis approaches to characterise single cell morphology and dynamics has increased. We present CellPhe, a pattern recognition toolkit for the unbiased characterisation of cellular phenotypes within time-lapse videos. CellPhe imports tracking information from multiple segmentation and tracking algorithms to provide automated cell phenotyping from different imaging modalities, including fluorescence. To maximise data quality for downstream analysis, our toolkit includes automated recognition and removal of erroneous cell boundaries induced by inaccurate tracking and segmentation. We provide an extensive list of features extracted from individual cell time series, with custom feature selection to identify variables that provide greatest discrimination for the analysis in question. Using ensemble classification for accurate prediction of cellular phenotype and clustering algorithms for the characterisation of heterogeneous subsets, we validate and prove adaptability using different cell types and experimental conditions.

A Comprehensive Investigation of the Impact of Class Overlap on Software Defect Prediction

Software Defect Prediction (SDP) is one of the most vital and cost-efficient operations to ensure the quality of software under developed. The performance of SDP heavily relies on the characteristics of experimental datasets (or say SDP datasets). However, there often exists the phenomenon of class overlap in the SDP datasets, i.e., defective modules and non-defective modules are similar in terms of values of metrics. Class overlap hinders the smooth performance as well as the use of SDP models. Even though efforts have been made to investigate the impact of overlapping instance removing techniques on the performance of SDP, many open issues are still challenging yet unknown. For example, 1) how to effectively identify the overlapping instances? 2) Whether is the phenomenon of class overlap universal in the SDP datasets? 3) What are the impacts of class overlap on the performance and interpretation of SDP models? Questions like these are very important but have not been fully explored yet. In this paper, we conduct an empirical study to comprehensively investigate the impact of class overlap on SDP. Specifically, we first propose an overlapping instances identification approach by analyzing the class distribution in the local neighborhood of a given instance. Based on the approach, we then investigate the impact of class overlap on the performance and the interpretation of seven representative SDP models. Finally, we investigate the impact of two common overlapping instance handling techniques (i.e., removing and separating techniques) on the performance of SDP models. Through an extensive case study on 230 datasets that span across industrial and open-source software projects, we observe that: i) 70.0% of SDP datasets contain overlapping instances; ii) different levels of class overlap have different impacts on the performance of SDP models. The class overlap ratio and the number of instances seriously affect the stability of the performance of SDP models; iii) class overlap affects the rank of the important feature list of SDP models, particularly the feature lists at the top 2 and top 3 ranks; IV) Class overlap handling techniques could statistically significantly improve the performance of SDP models trained on datasets with over 12.5% overlap ratios. Therefore, on the basis of these findings we suggest that future work in SDP should apply our proposed KNN method to: i) identify whether the overlap ratios of their defect datasets are greater than 12.5% before building SDP models; ii) remove the overlapping instances to find the more consistent guiding significance metrics; iii) combine RF classifier and class overlap handling techniques when reducing the efforts to review codes.

The Whale and the Microorganism: A Tale of a Classic Example and Linguistic Intuitions

A classic example of the arbitrary relation between the way a word sounds and its meaning is that microorganism is a very long word that refers to a very small entity, whereas whale is a very short word that refers to something very big. This example, originally presented in Hockett's list of language's design features, has been often cited over the years, not only by those discussing the arbitrary nature of language, but also by researchers of sound symbolism. While the two groups disagreed regarding the role of arbitrariness and sound symbolism in language, they both agreed there is a nonsound symbolic relation between word length and entity size in this case. This paper shows that the length of the words whale and microorganism in fact reflects a sound symbolic pattern. An analysis of >600 languages from >100 language families shows that languages use longer words to denote the concept small than they do to denote the concept big. The paper thus shows how explicit judgments might differ from implicit cognitive association and the problem of relying on these in sound symbolism research.

Identification of dynamic gene expression profiles during sequential vaccination with ChAdOx1/BNT162b2 using machine learning methods.

To date, COVID-19 remains a serious global public health problem. Vaccination against SARS-CoV-2 has been adopted by many countries as an effective coping strategy. The strength of the body's immune response in the face of viral infection correlates with the number of vaccinations and the duration of vaccination. In this study, we aimed to identify specific genes that may trigger and control the immune response to COVID-19 under different vaccination scenarios. A machine learning-based approach was designed to analyze the blood transcriptomes of 161 individuals who were classified into six groups according to the dose and timing of inoculations, including I-D0, I-D2-4, I-D7 (day 0, days 2-4, and day 7 after the first dose of ChAdOx1, respectively) and II-D0, II-D1-4, II-D7-10 (day 0, days 1-4, and days 7-10 after the second dose of BNT162b2, respectively). Each sample was represented by the expression levels of 26,364 genes. The first dose was ChAdOx1, whereas the second dose was mainly BNT162b2 (Only four individuals received a second dose of ChAdOx1). The groups were deemed as labels and genes were considered as features. Several machine learning algorithms were employed to analyze such classification problem. In detail, five feature ranking algorithms (Lasso, LightGBM, MCFS, mRMR, and PFI) were first applied to evaluate the importance of each gene feature, resulting in five feature lists. Then, the lists were put into incremental feature selection method with four classification algorithms to extract essential genes, classification rules and build optimal classifiers. The essential genes, namely, NRF2, RPRD1B, NEU3, SMC5, and TPX2, have been previously associated with immune response. This study also summarized expression rules that describe different vaccination scenarios to help determine the molecular mechanism of vaccine-induced antiviral immunity.

Classification of Micromobility Vehicles in Thermal-Infrared Images Based on Combined Image and Contour Features Using Neuromorphic Processing

Trends of environmental awareness, combined with a focus on personal fitness and health, motivate many people to switch from cars and public transport to micromobility solutions, namely bicycles, electric bicycles, cargo bikes, or scooters. To accommodate urban planning for these changes, cities and communities need to know how many micromobility vehicles are on the road. In a previous work, we proposed a concept for a compact, mobile, and energy-efficient system to classify and count micromobility vehicles utilizing uncooled long-wave infrared (LWIR) image sensors and a neuromorphic co-processor. In this work, we elaborate on this concept by focusing on the feature extraction process with the goal to increase the classification accuracy. We demonstrate that even with a reduced feature list compared with our early concept, we manage to increase the detection precision to more than 90%. This is achieved by reducing the images of 160 × 120 pixels to only 12 × 18 pixels and combining them with contour moments to a feature vector of only 247 bytes.

The Clothes Maketh the Cult

The Clothes Maketh the Cult

Establishing evidence to inform culturally competent mental health services: A mixed methods study protocol.

COVID-19 has exacerbated the significant and longstanding mental health inequalities for ethnic minorities, who were less likely to access mental health support in primary care but more likely to end up in crisis care compared to the majority ethnic group. Services were poorly offered and accessed to respond to the increased mental health challenges. To 1) establish evidence on which changes to mental health services provided in response to COVID-19 are beneficial for ethnic minorities who experience mental health difficulties in the North of England, and 2) to inform what and how culturally competent mental health services should be routinely provided. A mixed methods approach comprising 1) a rapid review to map services and models of care designed or adjusted for mental health during the pandemic, 2) an observational study of retrospective routine data to assess changes to mental health services and associated outcomes, 3) qualitative interviews to understand experiences of seeking care and factors associated with high-quality service provision, and 4) a Delphi study to establish consensus on key features of culturally competent services. From the selected areas in the North of England, adults from ethnic minorities who experience mental health difficulties will be identified from the primary, community and secondary care services and local ethnic minority communities. This study will identify ways to tackle health inequalities and contribute to mental health service recovery post pandemic by providing practice recommendations on equitable and effective services to ensure culturally competent and high-quality care. A list of services and features on what and how mental health services will be provided. Working with study collaborators and public and patient involvement partners, the study findings will be widely disseminated through presentations, conferences and publications and will inform the subsequent funding application for intervention development and evaluation.

Features Matching using Natural Language Processing

The feature matching is a basic step in matching different datasets. This article proposes shows a new hybrid model of a pretrained Natural Language Processing (NLP) based model called BERT used in parallel with a statistical model based on Jaccard similarity to measure the similarity between list of features from two different datasets. This reduces the time required to search for correlations or manually match each feature from one dataset to another.

Resolving operational paradox of sustainable supply chain: A decision framework approach

This study helps to reconcile the trade-offs between sustainable supply chain management (SSCM) and operational performance. Inspired by the paradox theory, the study advocates avoiding the contradictory elements and accepting the complementary elements between SSCM and Operational Performance. Accordingly, the study identifies various SSCM and operational performance elements followed by their evaluation to understand their complementary and contradictory nature. The study is conducted in two stages. First, the study constructs a list of SSCM features through exploratory factor analysis. Second, a unique decision framework of MACBETH (Measuring Attractiveness by Categorical-Based Evaluation Technique) and TODIM (Tomada de Decisión Inerativa Multicritero) methods is used to evaluate the SSCM features based on their impact on operational performance criteria. The proposed framework is validated in the context of the Indian Automobile Industry. The study results provide an empirically validated and prioritized list of SSCM features. In the list, the top-ranked features complement the operational performance criteria. In contrast, the lower-ranked features compromise the operational criteria, at least in the short term. Thus, this study reduces the skepticism around the adoption of SSCM by focusing on the top-ranked features of the list and avoiding the lower-ranked features during the early phases of SSCM adoption. Additionally, the study provides guidance to supply chain managers on achieving sustainability in a supply chain without compromising on its traditional goals. The results are of practical importance as supply chain managers may now choose to implement specific SSCM features that exhibit minimum negative impact on operational performance. Consequently, the study encourages an assertive adoption of SSCM even in developing countries.

Predictors of cognitive deterioration in subjective cognitive decline: evidence from longitudinal studies and implications for SCD-plus criteria

BackgroundSubjective cognitive decline (SCD) is an early manifestation of cognitive deterioration (CD) in some individuals. Therefore, it is worthwhile to conduct a systematic review and meta-analysis to summarise predictors of...

Creating A Reliable Digital Maternal Health Platform Using Maslow's Hierarchy and Nielsen Heuristics

This research seeks to build on previous efforts to serve the needs of expectant mothers in terms of providing effective digital healthcare. The first task in accomplishing this was to determine a list of basic features needed to completely satisfy all key performance indicators and needs of a maternal healthcare website. Once this was completed and each category of the established hierarchy had been addressed, the second task could be approached. This was to ensure the platform’s usability by designing sub features with the Nielsen Heuristics in mind. The result of these tasks is a complete framework for an all-in-one platform powerful enough to assist women through pregnancy. This research is intended for doctors, specifically those interested in the evolving virtual landscape, and other healthcare professionals. We believe these people have a significant stake in the development of virtual platforms and welcome their interest and criticism. The significance of this research lies in its future potential: the actualization of the platform. A sound framework for the feature set has been established, so what follows is the implementation, which will prove useful to doctors and expectant mothers alike.

Abstract P2-03-07: Multi-parametric algorithm integrating on-treatment Ki67 value and standard clinicopathological variables to predict risk of recurrences for women > 70 years old with early ER+HER2- tumours in POETIC trial

Abstract Background: Prognosis in older patients with breast cancer (BC) is worse compared to younger patients. No robust and specific tool to predict the risk of recurrence (TTR) for women aged 70 and over is likely due to the lack of representation of this group in the data from clinical trials. In the POETIC trial, of estrogen receptor-positive (ER+) and mainly human epidermal growth factor receptor 2 negative (HER2-) BC (88%), peri-operative aromatase inhibitor (POAI) did not improve treatment outcome, but patients with low baseline Ki67 value (Ki67B) or low POAI-induced Ki67 value (Ki67_2wk) had good outcome with standard of care therapy (usually adjuvant endocrine therapy (adjET) with the addition of chemotherapy as clinically indicated). In this study, we sought to develop a multi-parametric algorithm, named Ki67Cal, by integrating Ki67_2wk value with tumour characteristics to predict TTR for patients > 70 years old (yr) with early ER+HER2- BC treated with adjET only. Methods: Within POETIC, 39% (n=1744) of the randomised patients (n = 4480) were >70 years old. There were 813 patients aged > 70yr, with ER+HER2- BC, randomised to POAI treatment and treated with adjET only. A power calculation indicated that 811 such patients were sufficient to develop a prediction model that minimized overfitting, allowed up to 8 predictors, for predicting 5-years TTR with a median follow-up of 5.24 years and an overall event rate per 1000 person-years = 0.027, and provided an anticipated performance in terms of model fit R2 = 0.08 (Riley et al. BMJ 2020). A three-fold cross-validation approach was applied; an optimal list of features was selected in the training set (n = 538, events = 70); the agreement between expected and observed outcomes from the algorithm on the validation set (n = 275, events = 37) was evaluated by calibration plot. Multivariable Ridge Cox Regression model of significant parameters was built on the dataset merging training and validation datasets (n = 813) for precise estimates of the coefficients of parameters. A subset of post-POAI samples (n = 99) was gene expression profiled with Nanostring to allow pseudo-Oncotype, pseudo-EndoPredict, and RUO-Prosigna scores calculated (Buus et al. npj Breast cancer 2021). The risk groups classified by the Ki67Cal and gene-expression assays (GEP) were compared. Results: Within this cohort, the 5-year TTR was 34.5% (C.I. 24.9-47.9) for those with a high Ki67_2wk (>=10%) and 12.3% (C.I. 9.1-16.7) in those with a high Ki67B that was suppressed to Ki672wk < 10%. The significant features were Ki67_2wk, sampling type (core vs. excision) at surgery, and pathological variables (tumour size, grade, and nodal status) for the final Ki67Cal algorithm. Stratifying patients into five groups (quintiles) by Ki67Cal identified 60% of patients with TTR of < 5% at 5yrs, and 20% of patients with TTR of > 30% at 5yrs. As an exploratory analysis, the risk groups by Ki67Cal and GEP were compared (Table 1). To date, these assays are optimized to be used on untreated ER+HER2- samples; there were fairly good agreements between the high-risk group defined by Ki67Cal with pseudo-EndoPredict and RUO-Prosigna respectively, and low-risk groups by Ki67Cal with Prosigna probably because Prosigna scores are driven by proliferation score. Conclusion: The relatively poor outcome of patients >70yrs in POETIC emphasizes the need for prognostic tools that identify patients who may be treated with endocrine therapy alone or conversely should be considered for additional therapy. Ki67Cal provides a simple tool that identified very low-risk and high-risk patients in 80% of patients with ER+HER2- BC. Table 1: Comparison of the risk groups defined by Ki67Cal algorithm with the three commonly used gene-expression assays (pseudo-EndoPredict, pseudo-Oncotype and RUO-Prosigna) applied on the post-peri-operative aromatase inhibitor samples. Citation Format: Maggie Chon U Cheang, Monisha Dewan, Lucy Kilburn, Gabriele Morani, Lila Zabaglo, Kally Sidhu, Holly Tovey, Xixuan Zhu, Chris Holcombe, Anthony Skene, Ian Smith, John Robertson, Alistair Ring, Nicholas Turner, Judith Bliss, Mitch Dowsett. Multi-parametric algorithm integrating on-treatment Ki67 value and standard clinicopathological variables to predict risk of recurrences for women > 70 years old with early ER+HER2- tumours in POETIC trial [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P2-03-07.

Elicitation of domain knowledge for a machine learning model for paediatric critical illness in South Africa.

Delays in identification, resuscitation and referral have been identified as a preventable cause of avoidable severity of illness and mortality in South African children. To address this problem, a machine learning model to predict a compound outcome of death prior to discharge from hospital and/or admission to the PICU was developed. A key aspect of developing machine learning models is the integration of human knowledge in their development. The objective of this study is to describe how this domain knowledge was elicited, including the use of a documented literature search and Delphi procedure. A prospective mixed methodology development study was conducted that included qualitative aspects in the elicitation of domain knowledge, together with descriptive and analytical quantitative and machine learning methodologies. A single centre tertiary hospital providing acute paediatric services. Three paediatric intensivists, six specialist paediatricians and three specialist anaesthesiologists. None. The literature search identified 154 full-text articles reporting risk factors for mortality in hospitalised children. These factors were most commonly features of specific organ dysfunction. 89 of these publications studied children in lower- and middle-income countries. The Delphi procedure included 12 expert participants and was conducted over 3 rounds. Respondents identified a need to achieve a compromise between model performance, comprehensiveness and veracity and practicality of use. Participants achieved consensus on a range of clinical features associated with severe illness in children. No special investigations were considered for inclusion in the model except point-of-care capillary blood glucose testing. The results were integrated by the researcher and a final list of features was compiled. The elicitation of domain knowledge is important in effective machine learning applications. The documentation of this process enhances rigour in such models and should be reported in publications. A documented literature search, Delphi procedure and the integration of the domain knowledge of the researchers contributed to problem specification and selection of features prior to feature engineering, pre-processing and model development.

Early prediction of cardiovascular disease using artificial neural network

Abstract In recent years, there has been a noticeable rise in the mortality rate, and heart disease is a significant contributor to this trend. According to the California Association for Diabetes Investigation, by 2015, cardiovascular disease would be the primary cause of death in India, where 62 billion people live. Deficiencies in the heart’s ability to pump blood to and from the rest of the body are the leading cause of cardiovascular disease. The healthcare industry is a prime example of a sector poised to greatly benefit from the availability of massive amounts of data and analytical insights. Increasingly, it will be important to extract medical data to predict and treat the high fatality rate caused by heart attacks. Every day, humanity generates terabytes worth of data. Medical errors with dire effects can be avoided only with high-quality services. Hospitals can reduce the price of expensive clinical testing by using decision support systems. Hospitals in the modern-day use hospital information systems to keep track of patient records. The health care sector generates vast amounts of data, but little of it is really put to good use. It will be important to adopt a new strategy to reduce costs and make accurate predictions about heart disease. To determine which machine learning and deep learning approaches are most useful and accurate for predicting and classifying cardiac illnesses, this article reviews the existing literature on the topic and subsequently tries to detect the most probable factors leading to heart disease. This study introduces and models an artificial neural network methodology for identifying potential cardiovascular disease risk factors. In this study, we examine and present the various full and partial correlations among risk attributes. In addition, a number of risk variables are analysed to generate a predicted list of risk features most likely to result in cardiovascular disease.