Search results for: long-term recall

Authors: Khadeeja Munawar, Shamsul Haque

Abstract:

The claim that traumatic events influence the recalled memories and mental health has received mixed empirical support. This study examines the memories of a sample drawn from Pakistan, a country that has witnessed many life-changing socio-political events, wars, and natural disasters in 72 years of its history. A sample of 210 senior citizens (Mage = 64.35, SD = 6.33) was recruited from Pakistan. The aim was to investigate if participants retrieved more memories related to past traumatic events using a word-cueing technique. Each participant reported ten memories to ten neutral cue words. The results revealed that past traumatic events were not adversely affecting the memories and mental health of participants. When memories were plotted with respect to the ages at which the events happened, a pronounced bump at 11-20 years of age was seen. Memories within as well as outside of the bump were mostly positive. The multilevel logistic regression modelling showed that the memories recalled were personally important and played a role in enhancing resilience. The findings revealed that despite facing an array of ethnic, religious, political, economic, and social conflicts, the participants were resilient, recalled predominantly positive memories, and had intact mental health. The findings have clinical implications in Cognitive Behavioral Therapy (CBT). The patients can be made aware of their negative emotions, troublesome/traumatic memories, and the distorted thinking patterns and their memories can be restructured. The findings can also be used to teach Memory Specificity Training (MEST) by psycho-educating the patients around changes in memory functioning and enhancing the recall of memories, which are more specific, vivid, and filled with sensory details.

Keywords: cognitive behavioral therapy, memories, mental health, resilience, trauma

Procedia PDF Downloads 120

Authors: Danijela Vranješ, Nataša Vukajlović

Abstract:

Turning to learner-centered teaching instead of the teacher-centered approach brought a whole new perspective into the process of teaching and learning and set a new goal for improving the educational process itself. However, recently a tremendous decline in students’ performance on various standardized tests can be observed, above all on the PISA-test. The learner-centeredness on its own is not enough anymore: the students’ ability to think is deteriorating. Especially in foreign language learning, one can encounter a lot of learning by heart: whether it is grammar or vocabulary, teachers often seem to judge the students’ success merely on how well they can recall a specific word, phrase, or grammar rule, but they rarely aim to foster their ability to think. Convinced that foreign language teaching can do both, this research aims to discover how two different approaches to teaching foreign language foster the students’ ability to think as well as to what degree they help students get to the state-determined level of foreign language at the end of the semester as defined in the Common European Framework. For this purpose, two different curricula were developed: one is a traditional, learner-centered foreign language curriculum that aims at teaching the four competences as defined in the Common European Framework and serves as a control variable, whereas the second one has been enriched with various thinking routines and aims at teaching the foreign language as a means to communicate ideas and thoughts rather than reducing it to the four competences. Moreover, two types of tests were created for each approach, each based on the content taught during the semester. One aims to test the students’ competences as defined in the CER, and the other aims to test the ability of students to draw on the knowledge gained and come to their own conclusions based on the content taught during the semester. As it is an ongoing study, the results are yet to be interpreted.

Keywords: common european framework of reference, foreign language learning, foreign language teaching, testing and assignment

Procedia PDF Downloads 71

Authors: Harshit Mittal, Neeraj Garg

Abstract:

Hand symbol recognition is a pivotal component in the domain of computer vision, with far-reaching applications spanning sign language interpretation, human-computer interaction, and accessibility. This research paper discusses the approach with the integration of the Canny Edge algorithm and convolutional neural network. The significance of this study lies in its potential to enhance communication and accessibility for individuals with hearing impairments or those engaged in gesture-based interactions with technology. In the experiment mentioned, the data is manually collected by the authors from the webcam using Python codes, to increase the dataset augmentation, is applied to original images, which makes the model more compatible and advanced. Further, the dataset of about 6000 coloured images distributed equally in 5 classes (i.e., 1, 2, 3, 4, 5) are pre-processed first to gray images and then by the Canny Edge algorithm with threshold 1 and 2 as 150 each. After successful data building, this data is trained on the Convolutional Neural Network model, giving accuracy: 0.97834, precision: 0.97841, recall: 0.9783, and F1 score: 0.97832. For user purposes, a block of codes is built in Python to enable a window for hand symbol recognition. This research, at its core, seeks to advance the field of computer vision by providing an advanced perspective on hand sign recognition. By leveraging the capabilities of the Canny Edge algorithm and convolutional neural network, this study contributes to the ongoing efforts to create more accurate, efficient, and accessible solutions for individuals with diverse communication needs.

Keywords: hand symbol recognition, computer vision, Canny edge algorithm, convolutional neural network

Procedia PDF Downloads 33

Authors: Brindha Senthil Kumar, Payel Chakraborty, Senthil Kumar Nachimuthu, Arindam Maitra, Prem Nath

Abstract:

Gastric cancer is predominantly caused by demographic and diet factors as compared to other cancer types. The aim of the study is to predict Early Gastric Cancer (ECG) from diet and lifestyle factors using supervised machine learning algorithms. For this study, 160 healthy individual and 80 cases were selected who had been followed for 3 years (2016-2019), at Civil Hospital, Aizawl, Mizoram. A dataset containing 11 features that are core risk factors for the gastric cancer were extracted. Supervised machine algorithms: Logistic Regression, Naive Bayes, Support Vector Machine (SVM), Multilayer perceptron, and Random Forest were used to analyze the dataset using Python Jupyter Notebook Version 3. The obtained classified results had been evaluated using metrics parameters: minimum_false_positives, brier_score, accuracy, precision, recall, F1_score, and Receiver Operating Characteristics (ROC) curve. Data analysis results showed Naive Bayes - 88, 0.11; Random Forest - 83, 0.16; SVM - 77, 0.22; Logistic Regression - 75, 0.25 and Multilayer perceptron - 72, 0.27 with respect to accuracy and brier_score in percent. Naive Bayes algorithm out performs with very low false positive rates as well as brier_score and good accuracy. Naive Bayes algorithm classification results in predicting ECG showed very satisfactory results using only diet cum lifestyle factors which will be very helpful for the physicians to educate the patients and public, thereby mortality of gastric cancer can be reduced/avoided with this knowledge mining work.

Keywords: Early Gastric cancer, Machine Learning, Diet, Lifestyle Characteristics

Procedia PDF Downloads 117

Authors: William Cheng, Rini Sekartini

Abstract:

Background: Sleep is essential for children because it provides enhancement for the neural system activities that give physiologic effects for the body to support growth and development. One of the modifiable factors that relates with sleep is nutrition, which includes nutritional status, iron intake, and magnesium intake. Nutritional status represents the balance between nutritional intake and expenditure, while iron and magnesium are micronutrients that are related to sleep regulation. The aim of this study is to identify prevalence of sleep disturbance among Indonesian children and to evaluate its relation with aspect to nutrition. Methods : A cross-sectional study involving children aged 5 to 7-years-old in an urban primary health care between 2012 and 2013 was carried out. Related data includes anthropometric status, iron intake, and magnesium intake. Iron and magnesium intake was obtained by 24-hours food recall procedure. Sleep Disturbance Scale for Children (SDSC) was used as the diagnostic tool for sleep disturbance, with score under 39 indicating presence of problem. Results: Out of 128 school-aged children included in this study, 28 (23,1%) of them were found to have sleep disturbance. The majority of children had good nutritional status, with only 15,7% that were severely underweight or underweight, and 12,4% that were identified as stunted. On the contrary, 99 children (81,8%) were identified to have inadequate magnesium intake and 56 children (46,3%) with inadequate iron intake. Our analysis showed there was no significant relation between all of the nutritional status indicators and sleep disturbance (p>0,05%). Moreover, inadequate iron and magnesium intake also failed to prove significant relation with sleep disturbance in this population. Conclusion: Almost fourth of school-aged children in Indonesia were found to have sleep disturbance and further study are needed to overcome this problem. According to our finding, there is no correlation between nutritional status, iron intake, magnesium intake, and sleep disturbance.

Keywords: iron intake, magnesium intake, nutritional status, school-aged children, sleep disturbance

Procedia PDF Downloads 439

Authors: Yufeng Li, Yinan Ma, Jing Wu, Chengnian Long

Abstract:

Infrared object detection aims at separating small and dim targets from cluttered backgrounds, and its capabilities extend beyond the limits of visible light, making it invaluable in a wide range of applications, such as improving safety, security, efficiency, and functionality. However, existing methods are usually sensitive to the noise of the input infrared image, leading to a decrease in target detection accuracy and an increase in the false alarm rate in high-noise environments. To address this issue, an infrared small target detection algorithm called YOLO-IR is proposed in this paper to improve the robustness to high infrared noise. To address the problem that high noise significantly reduces the clarity and reliability of target features in infrared images, we design a soft-threshold coordinate attention mechanism to improve the model’s ability to extract target features and its robustness to noise. Since the noise may overwhelm the local details of the target, resulting in the loss of small target features during depth down-sampling, we propose a deep and shallow feature fusion neck to improve the detection accuracy. In addition, because the generalized Intersection over Union (IoU)-based loss functions may be sensitive to noise and lead to unstable training in high-noise environments, we introduce a Wasserstein-distance based loss function to improve the training of the model. The experimental results show that YOLO-IR achieves a 5.0% improvement in recall and a 6.6% improvement in the F1 score over the existing state-of-the-art model.

Keywords: infrared small target detection, high noise, robustness, soft-threshold coordinate attention, feature fusion

Procedia PDF Downloads 25

Authors: James Ormrod

Abstract:

This paper is concerned with volunteer livestock checking. Based on a pilot study consisting of ‘go-along’ interviews with livestock checkers, it argues that there are limitations to the insights that can be generated from approaches to ‘discourse analysis’ that would focus only on the verbalizable aspects of the practice. Volunteer livestock checking takes place across Europe as part of conservation projects aimed at maintaining particular habitats through the reintroduction of grazing animals. Volunteers are variously called ‘urban shepherds’, because these practices often take place on urban fringes, or ‘lookerers’, as their role is to make visual checks on the animals. Pilot research that took place on the South Downs (a chalk downland habitat on the South Coast of the UK) involved researchers accompanying volunteers as they checked on livestock. They were asked to give an account of what they were doing and then answer semi-structured interview questions. Participants drew on popular discourses on conservation and biodiversity, as framed by the local council who run the programme. They also framed their relationships to the animals in respect to the more formal limitations of their role as identified through the conservation programme. And yet these discourses, significant as they are, do not adequately explain why volunteers are drawn to, and emotionally invested in, lookering. The methodology employed allowed participants instead to gesture to features of the landscape and to recall memories, and for the researchers to see how volunteers interacted with the animals and the landscape in embodied and emotionally loaded ways. The paper argues that a psychosocial perspective that pays attention to the contradictions and tensions made visible through this methodology helps develop a fuller understanding of volunteer livestock checking as a social practice.

Keywords: conservation, human-animal relations, lookering, volunteering

Procedia PDF Downloads 105

Authors: Jungbin Lim, Bohee Kang, Heelim Lee, Sunjin Kim, GeumHee Choi, Jae-Suk Jung, Suk Ji

Abstract:

Nonsurgical periodontal therapies have, for several decades, been the basis of periodontal treatment concepts. The aim of this paper is to investigate the effectiveness of non-surgical periodontal therapy according to the severity of periodontitis disease. Methods: Retrospective data of patients who visited Department of periodontics in Ajou University Medical Center from 2016 to 2022 were collected. Among the patients, those who took full mouth examination of clinical parameters and non-surgical periodontal therapy were chosen for this study. Selected patients were divided into initial, moderate, and severe periodontitis based on severity and complexity of management (2018 World Workshop EFP/AAP consensus). Recall visits with clinical periodontal examination were scheduled for 1,2,3 months or 1,3,6 months after the treatment. The results were evaluated by recordings of mean probing pocket depth (mean PD), mean clinical attachment levels (mean CAL), bleeding on probing (BOP%), mean gingival index (mean GI), mean regression, mean sulcus bleeding index (mean SBI), mean plaque scores (mean PI). All statistical analyses were performed with R software, version 4.3.0. A level of significance, P<0.05, was considered to be statistically significant. Results: A total of 92 patients were included in this study. 15 patients were diagnosed as initial periodontitis, 14 moderate periodontitis, and 63 severe periodontitis. The all parameters except for mean recession decreased over time in all groups. The amount of mean PD decreased were the greatest in severe periodontitis group followed by moderate and initial, which was found to be statistically significant. The changes of mean PD were 0.15±0.05 mm, 0.37±0.06 mm, and 1.01±0.07 mm (initial, moderate, and severe, respectively, P<0.001). When comparing before and after treatment, the reductions in BOP(%), mean GI, mean SBI, and mean PI were statistically significant. Conclusion: All patients who received non-surgical periodontal therapy showed periodontal healing in terms of improvements in clinical parameters, and it was greater in the severe group.

Keywords: periodontology, clinical periodontology, oral treatment, comprehensive preventive dentistry, non-surgical periodontal therapy

Procedia PDF Downloads 41

Authors: Abeer A. Aljohani

Abstract:

COVID-19 virus spread has been one of the extreme pandemics across the globe. It is also referred to as coronavirus, which is a contagious disease that continuously mutates into numerous variants. Currently, the B.1.1.529 variant labeled as omicron is detected in South Africa. The huge spread of COVID-19 disease has affected several lives and has surged exceptional pressure on the healthcare systems worldwide. Also, everyday life and the global economy have been at stake. This research aims to predict COVID-19 disease in its initial stage to reduce the death count. Machine learning (ML) is nowadays used in almost every area. Numerous COVID-19 cases have produced a huge burden on the hospitals as well as health workers. To reduce this burden, this paper predicts COVID-19 disease is based on the symptoms and medical history of the patient. This research presents a unique architecture for COVID-19 detection using ML techniques integrated with feature dimensionality reduction. This paper uses a standard UCI dataset for predicting COVID-19 disease. This dataset comprises symptoms of 5434 patients. This paper also compares several supervised ML techniques to the presented architecture. The architecture has also utilized 10-fold cross validation process for generalization and the principal component analysis (PCA) technique for feature reduction. Standard parameters are used to evaluate the proposed architecture including F1-Score, precision, accuracy, recall, receiver operating characteristic (ROC), and area under curve (AUC). The results depict that decision tree, random forest, and neural networks outperform all other state-of-the-art ML techniques. This achieved result can help effectively in identifying COVID-19 infection cases.

Keywords: supervised machine learning, COVID-19 prediction, healthcare analytics, random forest, neural network

Procedia PDF Downloads 62

Authors: Ahmed A. A. Fadol, Guangji Tong, Wlaa Mohamed

Abstract:

This paper provides a comprehensive analysis of food insecurity in Sudan, focusing on consumption patterns and their implications, employing the Integrated Food Security Phase Classification (IPC) assessment framework. Years of conflict and economic instability have driven large segments of the population in Sudan into crisis levels of acute food insecurity according to the (IPC). A substantial number of people are estimated to currently face emergency conditions, with an additional sizeable portion categorized under less severe but still extreme hunger levels. In this study, we explore the multifaceted nature of food insecurity in Sudan, considering its historical, political, economic, and social dimensions. An analysis of consumption patterns and trends was conducted, taking into account cultural influences, dietary shifts, and demographic changes. Furthermore, we employ logistic regression and random forest analysis to identify significant independent variables influencing food security status in Sudan. Random forest clearly outperforms logistic regression in terms of area under curve (AUC), accuracy, precision and recall. Forward projections of the IPC for Sudan estimate that 15 million individuals are anticipated to face Crisis level (IPC Phase 3) or worse acute food insecurity conditions between October 2023 and February 2024. Of this, 60% are concentrated in Greater Darfur, Greater Kordofan, and Khartoum State, with Greater Darfur alone representing 29% of this total. These findings emphasize the urgent need for both short-term humanitarian aid and long-term strategies to address Sudan's deepening food insecurity crisis.

Keywords: food insecurity, consumption patterns, logistic regression, random forest analysis

Procedia PDF Downloads 32

Authors: Oumaima Khlifati, Khadija Baba

Abstract:

Pavement distress is the main factor responsible for the deterioration of road structure durability, damage vehicles, and driver comfort. Transportation agencies spend a high proportion of their funds on pavement monitoring and maintenance. The auscultation of pavement distress was based on the manual survey, which was extremely time consuming, labor intensive, and required domain expertise. Therefore, the automatic distress detection is needed to reduce the cost of manual inspection and avoid more serious damage by implementing the appropriate remediation actions at the right time. Inspired by recent deep learning applications, this paper proposes an algorithm for automatic road distress detection and classification using on the Deep Convolutional Neural Network (DCNN). In this study, the types of pavement distress are classified as transverse or longitudinal cracking, alligator, pothole, and intact pavement. The dataset used in this work is composed of public asphalt pavement images. In order to learn the structure of the different type of distress, the DCNN models are trained and tested as a multi-label classification task. In addition, to get the highest accuracy for our model, we adjust the structural optimization hyper parameters such as the number of convolutions and max pooling, filers, size of filters, loss functions, activation functions, and optimizer and fine-tuning hyper parameters that conclude batch size and learning rate. The optimization of the model is executed by checking all feasible combinations and selecting the best performing one. The model, after being optimized, performance metrics is calculated, which describe the training and validation accuracies, precision, recall, and F1 score.

Keywords: distress pavement, hyperparameters, automatic classification, deep learning

Procedia PDF Downloads 57

Authors: Nkechinyere Joy Olawuyi, Babajide Samuel Afolabi, Bola Ibitoye

Abstract:

This study collected a preprocessed dataset of chest radiographs and formulated a deep neural network model for detecting abnormalities. It also evaluated the performance of the formulated model and implemented a prototype of the formulated model. This was with the view to developing a deep neural network model to automatically classify abnormalities in chest radiographs. In order to achieve the overall purpose of this research, a large set of chest x-ray images were sourced for and collected from the CheXpert dataset, which is an online repository of annotated chest radiographs compiled by the Machine Learning Research Group, Stanford University. The chest radiographs were preprocessed into a format that can be fed into a deep neural network. The preprocessing techniques used were standardization and normalization. The classification problem was formulated as a multi-label binary classification model, which used convolutional neural network architecture to make a decision on whether an abnormality was present or not in the chest radiographs. The classification model was evaluated using specificity, sensitivity, and Area Under Curve (AUC) score as the parameter. A prototype of the classification model was implemented using Keras Open source deep learning framework in Python Programming Language. The AUC ROC curve of the model was able to classify Atelestasis, Support devices, Pleural effusion, Pneumonia, A normal CXR (no finding), Pneumothorax, and Consolidation. However, Lung opacity and Cardiomegaly had a probability of less than 0.5 and thus were classified as absent. Precision, recall, and F1 score values were 0.78; this implies that the number of False Positive and False Negative is the same, revealing some measure of label imbalance in the dataset. The study concluded that the developed model is sufficient to classify abnormalities present in chest radiographs into present or absent.

Keywords: transfer learning, convolutional neural network, radiograph, classification, multi-label

Procedia PDF Downloads 84

Authors: Carol Anne Hargreaves

Abstract:

A key issue in stock investment is how to select representative features for stock selection. The objective of this paper is to firstly determine whether an automated stock investment system, using machine learning techniques, may be used to identify a portfolio of growth stocks that are highly likely to provide returns better than the stock market index. The second objective is to identify the technical features that best characterize whether a stock’s price is likely to go up and to identify the most important factors and their contribution to predicting the likelihood of the stock price going up. Unsupervised machine learning techniques, such as cluster analysis, were applied to the stock data to identify a cluster of stocks that was likely to go up in price – portfolio 1. Next, the principal component analysis technique was used to select stocks that were rated high on component one and component two – portfolio 2. Thirdly, a supervised machine learning technique, the logistic regression method, was used to select stocks with a high probability of their price going up – portfolio 3. The predictive models were validated with metrics such as, sensitivity (recall), specificity and overall accuracy for all models. All accuracy measures were above 70%. All portfolios outperformed the market by more than eight times. The top three stocks were selected for each of the three stock portfolios and traded in the market for one month. After one month the return for each stock portfolio was computed and compared with the stock market index returns. The returns for all three stock portfolios was 23.87% for the principal component analysis stock portfolio, 11.65% for the logistic regression portfolio and 8.88% for the K-means cluster portfolio while the stock market performance was 0.38%. This study confirms that an automated stock investment system using machine learning techniques can identify top performing stock portfolios that outperform the stock market.

Keywords: machine learning, stock market trading, logistic regression, cluster analysis, factor analysis, decision trees, neural networks, automated stock investment system

Procedia PDF Downloads 127

Authors: Apostolos Kalatzis, Ashish Teotia, Vishnunarayan Girishan Prabhu, Laura Stanley

Abstract:

Emotions are functional behaviors influenced by thoughts, stimuli, and other factors that induce neurophysiological changes in the human body. Understanding and classifying emotions are challenging as individuals have varying perceptions of their environments. Therefore, it is crucial that there are publicly available databases and virtual reality (VR) based environments that have been scientifically validated for assessing emotional classification. This study utilized two commercially available VR applications (Guided Meditation VR™ and Richie’s Plank Experience™) to induce acute stress and calm state among participants. Subjective and objective measures were collected to create a validated multimodal dataset and classification scheme for affective state classification. Participants’ subjective measures included the use of the Self-Assessment Manikin, emotional cards and 9 point Visual Analogue Scale for perceived stress, collected using a Virtual Reality Assessment Tool developed by our team. Participants’ objective measures included Electrocardiogram and Respiration data that were collected from 25 participants (15 M, 10 F, Mean = 22.28  4.92). The features extracted from these data included heart rate variability components and respiration rate, both of which were used to train two machine learning models. Subjective responses validated the efficacy of the VR applications in eliciting the two desired affective states; for classifying the affective states, a logistic regression (LR) and a support vector machine (SVM) with a linear kernel algorithm were developed. The LR outperformed the SVM and achieved 93.8%, 96.2%, 93.8% leave one subject out cross-validation accuracy, precision and recall, respectively. The VR assessment tool and data collected in this study are publicly available for other researchers.

Keywords: affective computing, biosignals, machine learning, stress database

Procedia PDF Downloads 104

Authors: J. K. Pis-An

Abstract:

Objective: The purpose of the study was to examine the vulnerability of farmers to the effects of climate change, specifically the El Niño phenomenon was felt in the Philippines in 2009-2010. Methods: KAP Survey determines behavioral response to vulnerability to the effects of El Niño. Body Mass Index: Dietary Assessment using 24-hour food recall. Results: 75% of the respondents claimed that crop significantly decreased during drought. Indications that households of farmers are large where 51.6% are composed of 6-10 family members with 68% annual incomes below Php 100,00. Anthropometric assessment showed that the prevalence of Chronic Energy Deficiency Grade 1 among females 17% and 28.57% for low normal. While male body mass index result for chronic energy deficiency grade 1 10%, low normal 18.33% and and obese grade 1, 31.67%. Dietary assessment of macronutrient intake of carbohydrates, protein, and fat 31.6 % among respondents are below recommended amounts. Micronutrient deficiency of calcium, iron, vit. A, thiamine, riboflavin, niacin, and Vit. C. Conclusion: Majority of the rural populations are engaged into farming livelihood that makes up the backbone of their economic growth. Placing the current nutritional status of the farmers in the context of food security, there are reasons to believe that the status will go for worse if the extreme climatic conditions will once again prevail in the region. Farmers rely primarily on home grown crops for their food supply, a reduction in farm production during drought is expected to adversely affect dietary intake. The local government therefore institute programs to increase food resiliency and to prioritize health of the population as the moving force for productivity and development.

Keywords: world health organization, united nation framework convention on climate change, anthropometric, macronutrient, micronutrient

Procedia PDF Downloads 414

Authors: Widari

Abstract:

This study aims to determine the relationship of fast food consumption preferences with macro and micro nutrient adequacy students of SMP Negeri 5 Padang. This study used a cross sectional study conducted on 100 students of SMP Negeri 5 Padang. The variables studied were fast food preferences, nutrition adequacy macronutrients (carbohydrate, protein, fat, fiber) and micro nutrients (sodium, calcium, iron). Confounding factor in this study was the physical activity level because it was considered quite affecting food consumption of students. Data collected by using a questionnaire food recall as many as 2 x 24 hours to see the history of the respondents eat at school day and on holidays. Then, data processed using software Nutrisurvey and Microsoft Excel 2010. The analysis was performed on samples that have low and medium category on physical activity. The physical activity was not analyzed with another variable to see the strength of the relationship between independent and dependent variables. So that, do restrictions on physical activity variables in an attempt to get rid of confounding in design. Univariate and bivariate analyzes performed using SPSS 16.0 for Windows with Kolmogrov-Smirnov statistical tests, confidence level = 95% (α = 0,05). Results of univariate analysis showed that more than 70% of respondents liked fast food. On average, respondents were malnourished macro; malnourished fiber (100%), carbohydrates (72%), and protein (56%), whereas for fat, excess intake of the respondents (41%). Furthermor, many respondents who have micronutrient deficiencies; 98% for sodium, 96% for iron, and 91% for calcium. The results of the bivariate analysis showed no significant association between fast food consumption preferences with macro and micro nutrient adequacy (p > 0,05). This happens because in the fact not all students who have a preference for fast food actually eat them. To study better in the future, it is expected sampling really like and eat fast food in order to obtain better analysis results.

Keywords: fast food, nutritional adequacy, preferences, students

Procedia PDF Downloads 350

Authors: Peifang Hsieh

Abstract:

The prevention of child abuse has aroused serious concerns in Taiwan because of the disparity between the increasing amount of reported child abuse cases that doubled over the past decade and the scarcity of social workers. New Taipei city, with the most population in Taiwan and over 70% of its 4 million citizens are migrant families in which the needs of children can be easily neglected due to insufficient support from relatives and communities, sees urgency for a social support system, by preemptively identifying and outreaching high-risk families of child abuse, so as to offer timely assistance and preventive measure to safeguard the welfare of the children. Big data analysis is the inspiration. As it was clear that high-risk families of child abuse have certain characteristics in common, New Taipei city decides to consolidate detailed background information data from departments of social affairs, education, labor, and health (for example considering status of parents’ employment, health, and if they are imprisoned, fugitives or under substance abuse), to cross-reference for accurate and prompt identification of the high-risk families in need. 'The Service Center for High-Risk Families' (SCHF) was established to integrate data cross-departmentally. By utilizing the machine learning 'random forest method' to build a risk prediction model which can early detect families that may very likely to have child abuse occurrence, the SCHF marks high-risk families red, yellow, or green to indicate the urgency for intervention, so as to those families concerned can be provided timely services. The accuracy and recall rates of the above model were 80% and 65%. This prediction model can not only improve the child abuse prevention process by helping social workers differentiate the risk level of newly reported cases, which may further reduce their major workload significantly but also can be referenced for future policy-making.

Keywords: child abuse, high-risk families, big data analysis, risk prediction model

Procedia PDF Downloads 107

Authors: Omneya Messallam

Abstract:

Technology has been developed dramatically in most of the educational disciplines. For instance, digital rendering subject, which is being taught in both Interior and Architecture fields, is witnessing almost annually updated software versions. A lot of students and educators argued that there will be no need for manual rendering techniques to be learned. Therefore, the Interior Design Visual Presentation 1 course (ID133) has been chosen from the first level of the Interior Design (ID) undergraduate program, as it has been taught for six years continually. This time frame will facilitate sound observation and critical analysis of the use of appropriate teaching methodologies. Furthermore, the researcher believes in the high value of the manual rendering techniques. The course objectives are: to define the basic visual rendering principles, to recall theories and uses of various types of colours and hatches, to raise the learners’ awareness of the value of studying manual render techniques, and to prepare them to present their work professionally. The students are female Arab learners aged between 17 and 20. At the outset of the course, the majority of them demonstrated negative attitude, lacking both motivation and confidence in manual rendering skills. This paper is a reflective appraisal of deploying two student-centred teaching pedagogies which are: Project-based learning (PBL) and Outcome-based education (OBE) on ID133 students. This research aims of developing some teaching strategies to enhance the quality of teaching in this given course over an academic semester. The outcome of this research emphasized the positive influence of applying such educational methods on improving the quality of students’ manual rendering skills in terms of: materials, textiles, textures, lighting, and shade and shadow. Furthermore, it greatly motivated the students and raised the awareness of the importance of learning the manual rendering techniques.

Keywords: project-based learning, outcome-based education, visual presentation, manual render, personal competences

Procedia PDF Downloads 134

Authors: Malik Muhammad Arslan, Muneeb Ullah, Dai Shihan, Daniyal Haider, Xiaodong Yang

Abstract:

Chest X-rays are instrumental in the detection and monitoring of a wide array of diseases, including viral infections such as COVID-19, tuberculosis, pneumonia, lung cancer, and various cardiac and pulmonary conditions. To enhance the accuracy of diagnosis, artificial intelligence (AI) algorithms, particularly deep learning models like Convolutional Neural Networks (CNNs), are employed. However, these deep learning models demand a substantial and varied dataset to attain optimal precision. Generative Adversarial Networks (GANs) can be employed to create new data, thereby supplementing the existing dataset and enhancing the accuracy of deep learning models. Nevertheless, GANs have their limitations, such as issues related to stability, convergence, and the ability to distinguish between authentic and fabricated data. In order to overcome these challenges and advance the detection and classification of CXR normal and abnormal images, this study introduces a distinctive technique known as DCWGAN (Diverse Conditional Wasserstein GAN) for generating synthetic chest X-ray (CXR) images. The study evaluates the effectiveness of this Idiosyncratic DCWGAN technique using the ResNet50 model and compares its results with those obtained using the traditional GAN approach. The findings reveal that the ResNet50 model trained on the DCWGAN-generated dataset outperformed the model trained on the classic GAN-generated dataset. Specifically, the ResNet50 model utilizing DCWGAN synthetic images achieved impressive performance metrics with an accuracy of 0.961, precision of 0.955, recall of 0.970, and F1-Measure of 0.963. These results indicate the promising potential for the early detection of diseases in CXR images using this Inimitable approach.

Keywords: CNN, classification, deep learning, GAN, Resnet50

Procedia PDF Downloads 45

Authors: Nimot Opeyemi Wahab

Abstract:

Poor nutritional status during adolescence is a reflection of inadequate intake of nutrients. This can also be associated with a lack of consumption of diverse food. This study assessed the nutritional status and dietary diversity score (DDS) of in-school adolescents in Ibadan North, North East, and Ibadan South West Local Government Areas (LGA) of Oyo State, Nigeria. A cross-sectional study involving 3,510 in-school adolescents from the three LGA was conducted. Nutrient intake was measured using a validated 24-hour dietary recall, and the anthropometric measurement was also taken. Dietary diversity score (DDS) was assessed using the Individual Dietary Diversity Score (WDDS) of nine food groups. Participants were between 10-19years, and the mean age was 14.76±1.68, 15.32±1.77, and 15.45±1.62 in Ibadan North, Ibadan North East, and Ibadan South West, respectively. About 48% of the participants were male (47.9%), while 52.1% were female. BMI-for-age showed that 92.1%, 5.4%, 2.1%, and 0.4% of the participants were normal, underweight, overweight, and obese, respectively. The mean energy intake (143.193±695.98) of the female respondents was more than that of the male respondents (1406.86±767.41). The macronutrients intake (protein, carbohydrates, fiber, and fats) of the female participants was also found to be more than that of the male participants, with a non-significant difference of 0.336, 0.530, 0.234, and 0.069 (at p< 0.05). Out of all the vitamin intake, only vitamin C was found to be statistically different (p=0.038) at p<0.05 between the male and female respondents. Of all the mineral intake, only phosphorus showed a higher intake (575.20±362.12) among female respondents than the male respondents. The mean DDS of all participants was 4.59±0.939. The majority of the participants, 1183 (80.9%), were within the medium DDS category, 9.9% were low, while 1.5% were in the high category: of which males were 474 (71.5%) and females were 709 (88.6%). Participants from Ibadan North were 941(88.5%), and those from South West were 242(60.5%). A non-significant difference in the mean score of participants from the two locations (p=0.467) was also found. A negative correlation exists between DDS and BMI-for age (-0.11), DDS, and energy intake (-0.46) in Ibadan North and South West LGA. The nutritional status of in-school adolescents was normal, and DDS was within the medium category. Nutrition intervention regarding the consumption of diverse food is necessary among adolescents.

Keywords: nutritional status, dietary diversity, adolescents, nutrient intake

Procedia PDF Downloads 40

Authors: Michelle Hartanto, Risheka Suthantirakumar

Abstract:

Introduction Staff redeployment, increased numbers of acutely unwell patients requiring resuscitation decision-making conversations, visiting restrictions, and varying guidance regarding resuscitation for patients with COVID-19 disrupted clinicians’ management of resuscitation and escalation decision-making processes. While it was generally accepted that the COVID-19 pandemic disturbed numerous aspects of the Recommended Summary Plan for Emergency Care and Treatment (ReSPECT) process in the United Kingdom, a process which establishes a patient’s CPR status and treatment escalation plans, the impact of the pandemic on clinicians’ attitudes towards these resuscitation and decision-making conversations was unknown. This was the first study to examine the impact of the COVID-19 pandemic on clinicians’ knowledge, skills, and attitudes towards the ReSPECT process. Methods A cross-sectional survey of clinicians at one acute teaching hospital in the UK was conducted. A questionnaire with a defined five-point Likert scale was distributed and clinicians were asked to recall their pre-pandemic views on ReSPECT and report their current views at the time of survey distribution (May 2020, end of the first COVID-19 wave in the UK). Responses were received from 171 clinicians, and self-reported views before and during the pandemic were compared. Results Clinicians reported they found managing ReSPECT conversations more challenging during the pandemic, especially when conducted over the telephone with relatives, and they experienced an increase in negative emotions before, during, and after conducting ReSPECT conversations. Our findings identified that due to the pandemic there was now a need for clinicians to receive training and support in conducting resuscitation and escalation decision-making conversations over the telephone with relatives and managing these processes.

Keywords: cardiopulmonary resuscitation, COVID-19 pandemic, DNACPR discussion, education, recommended summary plan for emergency care and treatment, resuscitation order

Procedia PDF Downloads 79

Authors: Abhilash Sahu, Amit Priyadarshi

Abstract:

In this paper, we will discuss the existence of weak solutions of the Kirchhoff type boundary value problem on the Sierpinski gasket. Where S denotes the Sierpinski gasket in R² and S₀ is the intrinsic boundary of the Sierpinski gasket. M: R → R is a positive function and h: S × R → R is a suitable function which is a part of our main equation. ∆p denotes the p-Laplacian, where p > 1. First of all, we will define a weak solution for our problem and then we will show the existence of at least two solutions for the above problem under suitable conditions. There is no well-known concept of a generalized derivative of a function on a fractal domain. Recently, the notion of differential operators such as the Laplacian and the p-Laplacian on fractal domains has been defined. We recall the result first then we will address the above problem. In view of literature, Laplacian and p-Laplacian equations are studied extensively on regular domains (open connected domains) in contrast to fractal domains. In fractal domains, people have studied Laplacian equations more than p-Laplacian probably because in that case, the corresponding function space is reflexive and many minimax theorems which work for regular domains is applicable there which is not the case for the p-Laplacian. This motivates us to study equations involving p-Laplacian on the Sierpinski gasket. Problems on fractal domains lead to nonlinear models such as reaction-diffusion equations on fractals, problems on elastic fractal media and fluid flow through fractal regions etc. We have studied the above p-Laplacian equations on the Sierpinski gasket using fibering map technique on the Nehari manifold. Many authors have studied the Laplacian and p-Laplacian equations on regular domains using this Nehari manifold technique. In general Euler functional associated with such a problem is Frechet or Gateaux differentiable. So, a critical point becomes a solution to the problem. Also, the function space they consider is reflexive and hence we can extract a weakly convergent subsequence from a bounded sequence. But in our case neither the Euler functional is differentiable nor the function space is known to be reflexive. Overcoming these issues we are still able to prove the existence of at least two solutions of the given equation.

Keywords: Euler functional, p-Laplacian, p-energy, Sierpinski gasket, weak solution

Procedia PDF Downloads 211

Authors: Fangfei Li

Abstract:

Studies on teacher feedback have produced a wide range of findings in aspects of characteristics of good feedback, factors influencing the quality of feedback and teachers’ perspectives on teacher feedback. However, perspectives from students on how they perceive and respond to teacher feedback are still under scrutiny. Especially for Chinese overseas students who come from a feedback-sparse educational context in China, they might have different experiences when engaging with teacher feedback in the UK Higher Education. Therefore, the research aims to investigate and shed some new light on how Chinese students engage with teacher feedback in the UK higher education and how teacher feedback could enhance their learning. Research questions of this study are 1) What are Chinese overseas students’ perceptions of teacher feedback in courses of the UK higher education? 2) How do they respond to the teacher feedback they obtained? 3) What factors might influence their’ engagement with teacher feedback? Qualitative case studies of five Chinese postgraduate students in a UK university have been conducted by employing various types of interviews, such as background interviews, scenario-based interviews, stimulated recall interviews and retrospective interviews to address the research inquiries. Data collection lasted seven months, covering two phases – the pre-sessional language programme and the first semester of the Master’s degree programme. Research findings until now indicate that some factors, such as tutors’ handwriting, implicit instruction and value comments, influence students understanding and internalizing tutor feedback. Except for difficulties in understanding tutor feedback, students’ responses to tutor feedback are also influenced by quantity and quality of tutor-student communication, time constraints and trust to tutor feedback, etc. Findings also reveal that tutor feedback is able to improve students’ learning in aspects of promoting reflection on professional knowledge, promoting students’ communication with peers and tutors, increasing problem awareness and writing with the reader in mind. This paper will mainly introduce the research topic, the methodological procedure and research findings gained until now.

Keywords: Chinese students, students’ perceptions, teacher feedback, the UK higher education

Procedia PDF Downloads 233

Authors: Othmane Touri, Sanaa El Filali, El Habib Benlahmar

Abstract:

Natural Language Processing (NLP) has gained significant attention lately. It has proved its ability to analyze and extract insights from unstructured text data in various languages. It is found that one of the most popular NLP applications is sentiment analysis which aims to identify the sentiment expressed in a piece of text, such as positive, negative, or neutral, in multiple languages. While there are several multilingual NLP models available for sentiment analysis, there is a need to investigate their effectiveness in different contexts and applications. In this study, we aim to investigate the effectiveness of different multilingual NLP models for sentiment analysis on a dataset of online product reviews in multiple languages. The performance of several NLP models, including Google Cloud Natural Language API, Microsoft Azure Cognitive Services, Amazon Comprehend, Stanford CoreNLP, spaCy, and Hugging Face Transformers are being compared. The models based on several metrics, including accuracy, precision, recall, and F1 score, are being evaluated and compared to their performance across different categories of product reviews. In order to run the study, preprocessing of the dataset has been performed by cleaning and tokenizing the text data in multiple languages. Then training and testing each model has been applied using a cross-validation approach where randomly dividing the dataset into training and testing sets and repeating the process multiple times has been used. A grid search approach to optimize the hyperparameters of each model and select the best-performing model for each category of product reviews and language has been applied. The findings of this study provide insights into the effectiveness of different multilingual NLP models for Multilingual Sentiment Analysis and their suitability for different languages and applications. The strengths and limitations of each model were identified, and recommendations for selecting the most performant model based on the specific requirements of a project were provided. This study contributes to the advancement of research methods in multilingual NLP and provides a practical guide for researchers and practitioners in the field.

Keywords: NLP, multilingual, sentiment analysis, texts

Procedia PDF Downloads 58

Authors: Ermias A. Tegegn, Million Meshesha

Abstract:

Human Immunodeficiency Virus and Acquired Immunodeficiency Syndrome (HIV/AIDS) is a major cause of death for most African countries. Ethiopia is one of the seriously affected countries in sub Saharan Africa. Previously in Ethiopia, having HIV/AIDS was almost equivalent to a death sentence. With the introduction of Antiretroviral Therapy (ART), HIV/AIDS has become chronic, but manageable disease. The study focused on a data mining technique to predict future living status of HIV/AIDS patients at the time of drug regimen change when the patients become toxic to the currently taking ART drug combination. The data is taken from University of Gondar Hospital ART program database. Hybrid methodology is followed to explore the application of data mining on ART program dataset. Data cleaning, handling missing values and data transformation were used for preprocessing the data. WEKA 3.7.9 data mining tools, classification algorithms, and expertise are utilized as means to address the research problem. By using four different classification algorithms, (i.e., J48 Classifier, PART rule induction, Naïve Bayes and Neural network) and by adjusting their parameters thirty-two models were built on the pre-processed University of Gondar ART program dataset. The performances of the models were evaluated using the standard metrics of accuracy, precision, recall, and F-measure. The most effective model to predict the status of HIV patients with drug regimen substitution is pruned J48 decision tree with a classification accuracy of 98.01%. This study extracts interesting attributes such as Ever taking Cotrim, Ever taking TbRx, CD4 count, Age, Weight, and Gender so as to predict the status of drug regimen substitution. The outcome of this study can be used as an assistant tool for the clinician to help them make more appropriate drug regimen substitution. Future research directions are forwarded to come up with an applicable system in the area of the study.

Keywords: HIV drug regimen, data mining, hybrid methodology, predictive model

Procedia PDF Downloads 113

Authors: Sanchir Erdenebayar, Namuuntsetseg Oyunbaatar

Abstract:

Objectives: Peritoneal dialysis treatment is one of the important forms of kidney replacement therapy, and it has recently developed instantly in Mongolia for the past five years. Currently, more than 120 patients undergo peritoneal dialysis nationwide. These patients lack nutritional education, which predisposes them to protein deficiency and further impairs their quality of life. However, there is no study which is conducted among those about their dietary in Mongolia. Therefore, integrated nutrition information and educating them about dietary patterns to follow are urgently needed for PD patients. Methods: A cross-sectional study was carried out on 45 patients aged between 18 and 60 years who were undergoing CAPD at the biggest Medvic dialysis center in Ulaanbaatar. The knowledge of nutrition and food intake is assessed by interview based on a validated questionnaire prepared from KDIGO guidelines, semi-FFQ and a 24-hour dietary recall method. In addition, a biochemical blood test that includes total protein, albumin, calcium, phosphorus, potassium, and hemoglobin is used for an assessment of the patient’s current nutritional status. Results: Knowledge of nutritional status for CAPD was great, with 21.4% of patients and 78.65% having poor nutrition knowledge. The rate of mild to moderate malnutrition was 48.8% among research participants. Serum albumin was 38.4 ± 4.7 g/L, and total protein was 67.3±7.5g/l. Patients met 62.5± 26.5% of their daily intake nutritional requirement for calories and 72±40% of their nutritional requirement for protein. All patients’ energy intake was significantly /1328±304kcal/ lower than the energy requirement (2124±378kcal). Only 14.2% met the recommended dietary protein intake recommended to them of greater than 1.2 g/kg. Conclusions: As was established before, nutritional education has a vital positive impact on the health and nutritional status of peritoneal dialysis patients. The results of this study show that nutritional education programs are not enough adequate in peritoneal dialysis patients. There is a crucial priority to establish nutritional educational programs and guidelines for PD patients in Mongolia.

Keywords: renal diet, peritoneal dialysis, nutrition education, CKD diet

Procedia PDF Downloads 16

Authors: Edward Holupka, John Rossman, Tye Morancy, Joseph Aronovitz, Irving Kaplan

Abstract:

A common modality for the treatment of early stage prostate cancer is the implantation of radioactive seeds directly into the prostate. The radioactive seeds are positioned inside the prostate to achieve optimal radiation dose coverage to the prostate. These radioactive seeds are positioned inside the prostate using Transrectal ultrasound imaging. Once all of the planned seeds have been implanted, two dimensional transaxial transrectal ultrasound images separated by 2 mm are obtained through out the prostate, beginning at the base of the prostate up to and including the apex. A common deep neural network, called DetectNet was trained to automatically determine the position of the implanted radioactive seeds within the prostate under ultrasound imaging. The results of the training using 950 training ultrasound images and 90 validation ultrasound images. The commonly used metrics for successful training were used to evaluate the efficacy and accuracy of the trained deep neural network and resulted in an loss_bbox (train) = 0.00, loss_coverage (train) = 1.89e-8, loss_bbox (validation) = 11.84, loss_coverage (validation) = 9.70, mAP (validation) = 66.87%, precision (validation) = 81.07%, and a recall (validation) = 82.29%, where train and validation refers to the training image set and validation refers to the validation training set. On the hardware platform used, the training expended 12.8 seconds per epoch. The network was trained for over 10,000 epochs. In addition, the seed locations as determined by the Deep Neural Network were compared to the seed locations as determined by a commercial software based on a one to three months after implant CT. The Deep Learning approach was within \strikeout off\uuline off\uwave off2.29\uuline default\uwave default mm of the seed locations determined by the commercial software. The Deep Learning approach to the determination of radioactive seed locations is robust, accurate, and fast and well within spatial agreement with the gold standard of CT determined seed coordinates.

Keywords: prostate, deep neural network, seed implant, ultrasound

Procedia PDF Downloads 166

Authors: Mahmoud B. Rokaya

Abstract:

The study of opinion mining and sentiment analysis includes analysis of opinions, sentiments, evaluations, attitudes, and emotions. The rapid growth of social media, social networks, reviews, forum discussions, microblogs, and Twitter, leads to a parallel growth in the field of sentiment analysis. The field of sentiment analysis tries to develop effective tools to make it possible to capture the trends of people. There are two approaches in the field, lexicon-based and corpus-based methods. A lexicon-based method uses a sentiment lexicon which includes sentiment words and phrases with assigned numeric scores. These scores reveal if sentiment phrases are positive or negative, their intensity, and/or their emotional orientations. Creation of manual lexicons is hard. This brings the need for adaptive automated methods for generating a lexicon. The proposed method generates dynamic lexicons based on the corpus and then classifies text using these lexicons. In the proposed method, different approaches are combined to generate lexicons from text. The proposed method classifies the tweets into 5 classes instead of +ve or –ve classes. The sentiment classification problem is written as an optimization problem, finding optimum sentiment lexicons are the goal of the optimization process. The solution was produced based on mathematical programming approaches to find the best lexicon to classify texts. A genetic algorithm was written to find the optimal lexicon. Then, extraction of a meta-level feature was done based on the optimal lexicon. The experiments were conducted on several datasets. Results, in terms of accuracy, recall and F measure, outperformed the state-of-the-art methods proposed in the literature in some of the datasets. A better understanding of the Arabic language and culture of Arab Twitter users and sentiment orientation of words in different contexts can be achieved based on the sentiment lexicons proposed by the algorithm.

Keywords: social media, Twitter sentiment, sentiment analysis, lexicon, genetic algorithm, evolutionary computation

Procedia PDF Downloads 152

Authors: Archana Surendran, Kalpana C. A.

Abstract:

Vitamin D insufficiency is highly prevalent in women. Vitamin D bioavailability could be reduced in obesity due to increased sequestration by white adipose tissue. Increased sun exposure due to more frequent outdoor physical activity as well as a diet rich in vitamin D could be the common cause of both higher levels of 25(OH)D and a more favorable lipid profile. The study was conducted with the aim to assess the obesity status among selected working women in Coimbatore, determine their lifestyle and physical activity pattern, study their dietary intake, estimate the vitamin D and lipid profile of selected women and associate the relationship between Vitamin D and obesity among the selected women. A total of 100 working women (non pregnant, non lactating) working in IT sector, hotels and teaching staff were selected for the study. Anthropometric measurements and dietary recall were conducted for all. The women were further categorized as obese and non-obese based on their BMI. Fifteen obese and 15 non-obese women were selected and their fasting blood glucose level, serum Vitamin D and lipid profile were measured. Association between serum vitamin D, lipid profile, anthropometric measurements, food intake and sun exposure was correlated. Fifty six percent of women in the age group between 25-39 years and 44 percent of women in the age group between 40-45 years were obese. Waist and hip circumference of women in the age group between 40-45 years (89.7 and 107.4 cm) were higher than that of obese women in the age group between 25-39 years (88.6 and 102.8 cm). There were no women with sufficient vitamin D levels. In the age group between 40-45 years (obese women), serum Vitamin D was inversely proportional to waist-hip ratio and LDL cholesterol. There was an inverse relationship between body fat percentage and Total cholesterol with serum vitamin D among the women of the age group between 25-39 years. Consumption of milk and milk products were low among women. Intake of calcium was deficit among the women in both the age groups and showed a negative correlation. Sun exposure was less for all the women. Findings from the study revealed that obese women with a higher consumption of fat and less intake of calcium-rich foods have low serum Vitamin D levels than the non-obese women. Thus, it can be concluded that there is an association between Vitamin D status and obesity among adult women.

Keywords: obesity, sun exposure, vitamin D, women

Procedia PDF Downloads 107

Authors: Dylan Elliott, Katya Pertsova

Abstract:

Preposition errors are some of the most common errors created by L2 speakers. In addition, improving error correction and detection methods remains an open issue in the realm of Natural Language Processing (NLP). This research investigates whether the bidirectional encoder representations from the transformers model (BERT) have the potential to correct preposition errors accurately enough to be useful in error correction software. This research finds that BERT performs strongly when the scope of its error correction is limited to preposition choice. The researchers used an open-source BERT model and over three hundred thousand edited sentences from Wikipedia, tagged for part of speech, where only a preposition edit had occurred. To test BERT’s ability to detect errors, a technique known as multi-level masking was used to generate suggestions based on sentence context for every prepositional environment in the test data. These suggestions were compared with the original errors in the data and their known corrections to evaluate BERT’s performance. The suggestions were further analyzed to determine if BERT more often agreed with the judgements of the Wikipedia editors. Both the untrained and fined-tuned models were compared. Finetuning led to a greater rate of error-detection which significantly improved recall, but lowered precision due to an increase in false positives or falsely flagged errors. However, in most cases, these false positives were not errors in preposition usage but merely cases where more than one preposition was possible. Furthermore, when BERT correctly identified an error, the model largely agreed with the Wikipedia editors, suggesting that BERT’s ability to detect misused prepositions is better than previously believed. To evaluate to what extent BERT’s false positives were grammatical suggestions, we plan to do a further crowd-sourcing study to test the grammaticality of BERT’s suggested sentence corrections against native speakers’ judgments.

Keywords: BERT, grammatical error correction, preposition error detection, prepositions

Procedia PDF Downloads 116