Learning Approach Research Articles

Teaching and learning processes for pre-service mathematics teachers: The case of systems of equations

This study aims to explore the teaching and learning processes of the school mathematics course for pre-service mathematics teachers in the case of systems of equations. Types of teaching and learning approaches and a symbol sense perspective were used as theoretical frameworks for this study. To do so, we conducted a qualitative case study, involving 35 pre-service mathematics teachers, in the form of classroom observations in one of the state universities in Bandung, Indonesia. The observations included the teaching and learning processes, a written formative test, and interviews. The results showed that the teaching and learning processes are implemented by using a deductive approach,<i> </i>and the written student work from the formative test and interviews revealed various solution strategies. From the symbol sense perspective, we conclude that the teaching and learning processes need to be enhanced to promote pre-service teachers’ procedural skills and conceptual understanding in a balanced manner.

Evaluating machine learning models for precipitation prediction in Casablanca City

Accurate precipitation forecasting is a vital task for many domains, such as agriculture, water management, flood prevention, and crop yield estimation. The use of machine learning (ML) approaches has improved precipitation forecasting accuracy, exhibiting promising results in capturing the intricate connections between various meteorological variables and precipitation patterns. However, given the vast array of available ML models, a comparative analysis is imperative for identifying the most effective models for precipitation prediction. This study aims to examine the capacities of ML algorithms to forecast precipitation based on weather data for the city of Casablanca, Morocco, which faces challenges in water management and climate change adaptation. Eight different ML models’ performances are compared: linear regression, polynomial regression, K-nearest neighbors (KNN), support vector machine (SVM), decision tree (DT), random forest (RF), XGBoost, and an ensemble learning model. These models are evaluated based on their mean absolute error (MAE), mean squared error (MSE), and R-squared (R2 ) value to determine their effectiveness. The study showcases the potential of ML models in predicting precipitation by utilizing meteorological parameters such as temperature, humidity, wind speed, and pressure.

Effectiveness of VGG19 in deep learning for brain tumor detection

Image processing in the diagnosis of disease is one of the jobs that is currently developing in the world of health. Diagnosis is carried out by utilizing the role of image processing to provide a level of accuracy in diagnosis results and provide efficiency to medical personnel. This research aims to develop a brain tumor object detection process using a deep learning (DL) approach to magnetic resonance images (MRI) images. This development was carried out to optimize the brain tumor diagnosis process by playing the role of the image extraction process. This research dataset was sourced from the M. Djamil Padang Provincial General Hospital with a total of 3370 MRI images. The results of this work report show that DL performance is capable of carrying out the detection process automatically with an accuracy level of 97,83%. The results of the development of the extraction process can work effectively in ensuring brain tumor objects are precise and accurate. Overall, this research can make a major contribution to maximizing the diagnosis process and assisting medical personnel in the early treatment of brain tumor patients.

Identification and validation of a machine learning model of complete response to radiation in rectal cancer reveals immune infiltrate and TGFβ as key predictors

It is uncertain which biological features underpin the response of rectal cancer (RC) to radiotherapy. No biomarker is currently in clinical use to select patients for treatment modifications. We identified two cohorts of patients (total N=249) with RC treated with neoadjuvant radiotherapy (45Gy/25) plus fluoropyrimidine. This discovery set included 57 cases with pathological complete response (pCR) to chemoradiotherapy (23%). Pre-treatment cancer biopsies were assessed using transcriptome-wide mRNA expression and targeted DNA sequencing for copy number and driver mutations. Biological candidate and machine learning (ML) approaches were used to identify predictors of pCR to radiotherapy independent of tumour stage. Findings were assessed in 107 cases from an independent validation set (GSE87211). Three gene expression sets showed significant independent associations with pCR: Fibroblast-TGFβ Response Signature (F-TBRS) with radioresistance; and cytotoxic lymphocyte (CL) expression signature and consensus molecular subtype CMS1 with radiosensitivity. These associations were replicated in the validation cohort. In parallel, a gradient boosting machine model comprising the expression of 33 genes generated in the discovery cohort showed high performance in GSE87211 with 90% sensitivity, 86% specificity. Biological and ML signatures indicated similar mechanisms underlying radiation response, and showed better AUC and p-values than published transcriptomic signatures of radiation response in RC. RCs responding completely to chemoradiotherapy (CRT) have biological characteristics of immune response and absence of immune inhibitory TGFβ signalling. These tumours may be identified with a potential biomarker based on a 33 gene expression signature. This could help select patients likely to respond to treatment with a primary radiotherapy approach as for anal cancer. Conversely, those with predicted radioresistance may be candidates for clinical trials evaluating addition of immune-oncology agents and stromal TGFβ signalling inhibition. The Stratification in Colorectal Cancer Consortium (S:CORT) was funded by the Medical Research Council and Cancer Research UK (MR/M016587/1).

Assessing electromagnetic field exposure levels in multi-active reconfigurable intelligent surface assisted 5G network

As 5G mobile networks continue to proliferate in dense urban environments, it becomes increasingly important to understand and mitigate excessive electromagnetic field (EMF) exposure. This study investigates how the downlink EMF exposure levels of 5G millimeter wave (mm-wave) mobile networks are influenced by the integration of multi-active reconfigurable intelligent surfaces (RISs), using a ray-tracing approach. Our research employs a comprehensive two-step methodology: Firstly, we introduce a new RIS-assisted 5G mm-wave network planning technique. This technique leverages a machine learning (ML) approach for the classification of multi-RIS clusters. The primary goal is to optimize coverage while minimizing the number of required RIS deployments. This is achieved by strategically placing RISs based on the ML classification, ultimately aiming to enhance network efficiency. Secondly, we conducted a thorough comparative analysis, evaluating the impact of both passive and active RISs on EMF exposure level throughout a dense urban environment. Passive RIS and active RIS differ in their adaptability to changing network conditions. The result shows that the influence of multi-active RISs on EMF exposure is significant (about 7.5 times higher) compared to passive RISs.

Applying convolutional neural network and Nadam optimization in flower classification

Flowers have a variety of shapes, colors and structures, the images of which need to be classified using guided learning techniques. Several studies classify flowers using machine learning, but their accuracy performance is not good. The thing is, the flowers come in a variety of colors that can sometimes look similar to the background. Therefore, this study aims to classify flowers using a convolutional neural network (CNN) and measure its performance. The method used is mixed methods by collecting existing data from previous studies and connecting it with the realities in the field. The Kozłowski and Steinbrener models were used, while the image data was obtained from the Oxford17 and Oxford102 dataset with 17 and 102 flower types, respectively. The results show 60% and 84% accuracy of CNN using the scratch and transfer learning approach for the Oxford17 dataset. The Oxford102 dataset shows 42% and 64%, respectively, with CNN from baseline and transfer learning.

Major depressive disorder: early detection using deep learning and pupil diameter

Major depressive disorder stands as a highly prevalent mental disorder on a global scale. Detecting depression at its early stages holds paramount importance for effective treatment. However, due to the coexistence of depression with other conditions and the subjective nature of diagnosis, early identification poses a significant challenge. In recent times, machine learning techniques have emerged as valuable tools for the development of automated depression estimation systems, aiding in the diagnostic process. In this particular study, a deep learning approach utilizing pupil diameter was employed to distinguish between individuals diagnosed with depression and those who are considered mentally healthy. Pupillometric recordings were collected from a total of 58 individuals, comprising 29 healthy individuals and 29 individuals diagnosed with depression. Pupil size was recorded every 4 ms. The performance of three pretrained convolutional neural networks (GoogLeNet, SqueezeNet, and AlexNet) was evaluated for depression classification using the pupil size data. The highest accuracy of 98.28% was obtained with AlexNet. This finding highlights the potential of utilizing pupil diameter as a reliable indicator for objectively measuring depression.

Insights of machine learning-based threat identification schemes in advanced network system

An advanced network system (ANS) is characterized by extensive communication features that can support a sophisticated collaborative network structure. This is essential to hosting various forms of upcoming modernized and innovative applications. Security is one of the rising concerns associated with ANS deployment. It is also noted that machine learning is one of the preferred cost-effective ways to optimize the security strength and address various ongoing security problems in ANS; however, it is still unknown about its overall effectivity scale. Hence, this paper contributes to a systematic review of existing variants of machine learning approaches to deal with threat identification in ANS. As ANS is a generalized form, this discussion considers the impact of existing machine learning approaches on its practical use cases. The paper also contributes towards critical gap analysis and highlights the study's potential learning outcome.

Hospital quality classification based on quality indicator data during the COVID-19 pandemic

This research aim is to propose a machine learning approach to automatically evaluate or categories hospital quality status using quality indicator data. This research was divided into six stages: data collection, pre-processing, feature engineering, data training, data testing, and evaluation. In 2020, we collected 5,542 data values for quality indicators from 658 Indonesian hospitals. However, we analyzed data from only 275 hospitals due to inadequate submission. We employed methods of machine learning such as decision tree (DT), gaussian naïve Bayes (GNB), logistic regression (LR), k-nearest neighbors (KNN), support vector machine (SVM), linear discriminant analysis (LDA) and neural network (NN) for research archive purposes. Logistic regression achieved a 70% accuracy rate, SVM a 68% accuracy rate, and neural network a 59.34% of accuracy. Moreover, K-nearest neighbors achieved a 54% of accuracy and decision tree a 41% accuracy. Gaussian-NB achieved a 32% accuracy rate. The linear discriminant analysis achieved the highest accuracy with 71%. It can be concluded that linear discriminant analysis is the algorithm suitable for hospital quality data in this research.

Tissue Classification After Bone-Anchored Hearing Implant Surgery: A Machine Learning Approach to Monitoring Skin Response.

Here, we aim to 1) expand the available evidence for the use of machine learning techniques for soft tissue classification after BCD surgery and 2) discuss the implications of such approaches toward the development of classification applications to aid in tissue monitoring. The application of machine learning techniques in the soft tissue literature has become a large field of study. One of the most commonly reported outcomes after percutaneous bone-conduction device (BCD) surgery is soft tissue health. Unfortunately, the classification of tissue around the abutment as healthy versus not healthy is a subjective process, even though such decisions can have implications for treatment (i.e., topical steroid versus surgical revision) and resources (e.g., clinician time). We built and tested a convolutional neural network (CNN) model for the classification of tissues that were rated as "green" (i.e., healthy), "yellow" (i.e., unhealthy minor), and "red" (i.e., unhealthy severe). Representative image samples were gathered from a regional bone-conduction amplification site (N = 398; 181 samples of green; 144 samples of yellow; 73 samples of red). The image samples were cropped, zoomed, and normalized. Feature extraction was then implemented and used as the input to train an advanced CNN model. Accuracy of image classification for the healthy ("green") versus not healthy ("yellow" and "red") model was approximately 87%. Accuracy of image classification for the unhealthy ("yellow") versus unhealthy ("red") model was approximately 94%. Monitoring tissue health is an ongoing challenge for BCD users and their clinicians not trained in soft tissue management (e.g., audiologists). If machine learning can aid in the classification of tissue health, this would have significant implications for stakeholders. Here we discuss how machine learning can be applied to tissue classification as a potential technological aid in the coming years.

Advances in Cerebral Palsy Treatment

Background: Cerebral palsy is a complex neurodevelopmental disorder with various etiological factors and treatment options. This narrative review aimed to summarize the causes of cerebral palsy, identify areas needing additional research in treatment approaches, and highlight areas requiring further investigation. In order to provide a thorough overview of management techniques to lessen the effects of the illness and its consequences, this review has drawn data from a number of studies. Introduction: Prematurity increases the risk of brain damage during the developing stage and accounts for a sizable fraction of cerebral palsy cases. In a sizable portion of cases, maternal diabetes and hypertension are listed as the main causes. Damage to the brain tissue results from hypoxic-ischemic injuries sustained during pregnancy that upset the equilibrium of oxidants and antioxidants. To alter the oxidative stress pathway and ease related issues, pharmacological treatments, such as therapeutic hypothermia, free radical inhibition therapy, and mitochondrial therapy, have been proposed. Therapeutic strategies, such as physiotherapy, occupational therapy, speech therapy, and surgical interventions, have added quality to the lives of the children. Some of the most recent developments in this area include the development of biomarkers for muscle activity detection, machine learning to predict the types of cerebral palsy and abnormal movements, disease prediction with eye images, wireless inertia measuring unit for spasticity detection, computerbased video analysis of typical and atypical infants, identification of intellectual disabilities with algorithms, and deep learning methods for predicting cerebral palsy. Methods: This narrative review is based on a careful analysis of numerous researches conducted on cerebral palsy, which have served as the basis for statistical distribution. It reviews the causes of cerebral palsy, available treatments, and ongoing research with the goal of providing physicians and researchers in the field with useful information. The objectives, study questions, inclusion criteria, and search approach have all been outlined in a thorough protocol. To find pertinent research published up to September 2021, a literature search was carried out using electronic databases, including Google Scholar, PubMed, Cochrane Library, Scopus, and Web of Science. A combination of pertinent keywords, such as "cerebral palsy," "management," "technology," "wearable technology," "prematurity," and "artificial intelligence," has been used in the search approach. Results: Recent advances in the field include the discovery of biomarkers for the detection of muscle activity, machine learning algorithms to predict the types of cerebral palsy and abnormal movements, disease prediction using eye images, wireless inertia measuring units for the detection of spasticity, computer-based video analysis for the detection of atypical infants, and algorithms to identify intellectual disabilities. Additionally, employing technologies, like virtual reality systems, electrical stimulators, activity trackers, machine learning, and deep learning approaches, has shown promise in evaluating, diagnosing, and predicting treatment outcomes linked to gait, upper limb, and lower limb function. Conclusion: Future research should examine the clinical application of nanomedicine, stem cell therapy, and cutting-edge therapeutic strategies to prevent hypoxic-ischemic damage in the developing brain. Additionally, research is required to effectively assist children with severe speech difficulties using alternate communication modalities and cutting-edge computational tools. The outcomes for people with cerebral palsy can be improved by combining interdisciplinary efforts with cutting-edge technological interventions.

Data-Driven Active Disturbance Rejection Control of Plant-Protection Unmanned Ground Vehicle Prototype: A Fuzzy Indirect Iterative Learning Approach

Data-Driven Active Disturbance Rejection Control of Plant-Protection Unmanned Ground Vehicle Prototype: A Fuzzy Indirect Iterative Learning Approach

Prediction of global ionospheric TEC using attention based bidirectional long short-term memory and gated recurrent unit

An accurate prediction of ionospheric total electron content (TEC) at the primary stage is essential for applications related to global navigation satellite systems (GNSS) under varying weather conditions. The previous TEC prediction schemes contribute for each time step that increases the prediction time. The eye contact phenomenon establishes a metaphorical connection which intends to capture and emphasize the attention worthy elements in a sequence. This research introduces a deep learning approach which is a combination of attention-based bidirectional long short-term memory and gated recurrent unit (Bi-LSTM GRU) to predict TEC in the ionosphere. Bidirectional LSTM is the better option for achieving durability when combined with a gated recurrent unit (GRU) to predict TEC in the ionosphere. The proposed approach is evaluated with the existing LSTM approach for root mean square error (RMSE) during training and validation. The RMSE while predicting the global ionospheric delay using the existing LSTM for 20 epochs is seen to be 0.004, whereas the existing approach achieves a training error of 0.003.

From Traditional Typing to Intelligent Insights: A Narrative Review of Directions Toward Targeted Therapies in Delirium.

Delirium is a heterogeneous syndrome characterized by an acute change in level of consciousness that is associated with inattention and disorganized thinking. Delirium affects most critically ill patients and is associated with poor patient-oriented outcomes such as increased mortality, longer ICU and hospital length of stay, and worse long-term cognitive outcomes. The concept of delirium and its subtypes has existed since nearly the beginning of recorded medical literature, yet robust therapies have yet to be identified. Analogous to other critical illness syndromes, we suspect the lack of identified therapies stems from patient heterogeneity and prior subtyping efforts that do not capture the underlying etiology of delirium. The time has come to leverage machine learning approaches, such as supervised and unsupervised clustering, to identify clinical and pathophysiological distinct clusters of delirium that will likely respond differently to various interventions. We use sedation in the ICU as an example of how precision therapies can be applied to critically ill patients, highlighting the fact that while for some patients a sedative drug may cause delirium, in another cohort sedation is the specific treatment. Finally, we conclude with a proposition to move away from the term delirium, and rather focus on the treatable traits that may allow precision therapies to be tested.

Machine Learning Approaches to Predict 30-Day Mortality Following PCI in an Australian Population

Machine Learning Approaches to Predict 30-Day Mortality Following PCI in an Australian Population

Recent advancements in postpartum depression prediction through machine learning approaches: a systematic review

Postpartum depression (PPD) is a significant mental health concern affecting mothers worldwide, irrespective of demographic factors. Detecting and managing PPD at an early stage is crucial for effective intervention. In the context of mental health, intelligent predictive models based on machine learning (ML) have emerged as valuable tools. However, there remains a relative scarcity of research specifically targeting postpartum mental health due to several prominent factors that collectively impede the widespread adoption and practical implementation of ML in the field of PPD. This paper provides an updated overview of ML approaches for PPD prediction. A systematic search across IEEE Xplore, PubMed, Science Direct, and Scopus yielded 1,074 relevant articles. The performance of ML algorithms varies depending on the dataset and the problem being addressed. Notably, the findings reveal that the random forest (RF) algorithm consistently demonstrates the highest predictive accuracy, followed by support vector machine (SVM), logistic regression (LR), XGBoost, and AdaBoost. The development of advanced data techniques in PPD has encouraged interdisciplinary collaboration between researchers in psychiatry and computer science that holds great potential for refining the accuracy and reliability of PPD predictive models, ultimately resulting in improved outcomes for mothers and their families through early detection, intervention, and support.

Citrus leaf disease detection through deep learning approach

The majority of people in the world directly or indirectly depend on agriculture. Plant diseases are a significant threat to agricultural production and food security. Due to its high nutritional value, citrus fruit is one of the most abundant fruits in the world. However, different diseases are responsible for degraded citrus production as well as financial losses to the farmers. Traditionally, visual observation by experts has been attended to diagnose plant diseases. Usually, plant leaf disease recognition methods mainly rely on expert experiences to manually extract the colour, composition, and other features of diseased leaf images. Black spot, greening, canker, and melanoses are four common citrus leaf diseases. Rapid and accurate diagnosis of these diseases is a demand of time. Deep learning is a promising solution to these problems. There are different types of deep learning architecture like ImageNet, GoogleNet, VGG16, ResNet50, and InceptionV3, which show promising results in different object detection. Though most of these benchmark models give almost similar accuracy. However, this paper uses two deep learning models to find the better ones for the detection of citrus leaf disease detection. Hence, InceptionV3 outperforms VGG16 in terms of accuracy.

Predicting within-city spatiotemporal variations in daily median outdoor ultrafine particle number concentrations and size in Montreal and Toronto, Canada.

Epidemiological evidence suggests that long-term exposure to outdoor ultrafine particles (UFPs, <0.1 μm) may have important human health impacts. However, less is known about the acute health impacts of these pollutants as few models are available to estimate daily within-city spatiotemporal variations in outdoor UFPs. Several machine learning approaches (i.e., generalized additive models, random forest models, and extreme gradient boosting) were used to predict daily spatiotemporal variations in outdoor UFPs (number concentration and size) across Montreal and Toronto, Canada using a large database of mobile monitoring measurements. Separate models were developed for each city and all models were evaluated using a 10-fold cross-validation procedure. In total, our models were based on measurements from 12,705 road segments in Montreal and 10,929 road segments in Toronto. Daily median outdoor UFP number concentrations varied substantially across both cities with 1st-99th percentiles ranging from 1389 to 181,672 in Montreal and 2472 to 118,544 in Toronto. Outdoor UFP size tended to be smaller in Montreal (mean [SD]: 34 nm [15]) than in Toronto (mean [SD]: 44 nm [25]). Extreme gradient boosting models performed best and explained the majority of spatiotemporal variations in outdoor UFP number concentrations (Montreal, R 2: 0.727; Toronto, R 2: 0.723) and UFP size (Montreal, R 2: 0.823; Toronto, R 2: 0.898) with slopes close to one and intercepts close to zero for relationships between measured and predicted values. These new models will be applied in future epidemiological studies examining the acute health impacts of outdoor UFPs in Canada's two largest cities.

Current Computational Methods for Protein-peptide Complex Structure Prediction.

Peptide-mediated protein-protein interactions (PPIs) play an important role in various biological processes. The development of peptide-based drugs to modulate PPIs has attracted increasing attention due to the advantages of high specificity and low toxicity. In the development of peptide-based drugs, one of the most important steps is to determine the interaction details between the peptide and the target protein. In addition to experimental methods, recently developed computational methods provide a cost-effective way for studying protein-peptide interactions. In this article, we carefully reviewed recently developed protein-peptide docking methods, which were classified into three groups: template-based docking, template-free docking, and hybrid method. Then, we presented available benchmarking sets and evaluation metrics for assessing protein-peptide docking performance. Furthermore, we discussed the use of molecular dynamics simulations, as well as deep learning approaches in protein-peptide complex prediction.

Optimizing credit card fraud detection: a deep learning approach to imbalanced datasets

Imbalanced datasets pose a significant challenge in credit card fraud detection, hindering the training effectiveness of models due to the scarcity of fraudulent cases. This study addresses the critical problem of data imbalance through an in-depth exploration of techniques, including cross-entropy loss minimization, weighted optimization, and synthetic minority oversampling technique-based resampling, coupled with deep neural networks (DNNs). The urgent need to combat class imbalances in credit card fraud datasets is underscored, emphasizing the creation of reliable detection models. The research method delves into the application of DNNs, strategically optimizing and resampling the dataset to enhance model performance. The study employs a dataset from October 2018, containing 284,807 transactions, with a mere 492 classified as fraudulent. Various resampling techniques, such as undersampling and SMOTE oversampling, are evaluated alongside weighted optimization. The results showcase the effectiveness of SMOTE oversampling, achieving an accuracy of 99.83% without any false negatives. The study concludes by advocating for flexible strategies, integrating cutting-edge machine learning methods, and developing adaptive defenses to safeguard against emerging financial risks in credit card fraud detection.