Keyword Matching Research Articles

Using Textual Analysis to Examine Student Engagement in Online Undergraduate Science Education.

This study investigates the relationship between student engagement and peer review performance in an online data science course. Employing a mixed-methods approach, we analyzed quantitative data on peer review metrics (word count, sentiment, and complexity) and qualitative insights from student comments from 2019 to 2021. Our findings indicate a positive correlation between higher student grades and more detailed, positive peer reviews, as measured by word count, sentiment, and keyword matches. While the overall changes in engagement metrics were not statistically significant, trends suggest increased engagement over time. These findings contribute to a growing body of research on the effectiveness of peer review in online learning environments and highlight its potential to foster student development in data science. Future research should address limitations such as sample size and explore broader applications across diverse online courses..

B-101 Assessing the Performance of Cerebrospinal Fluid Kappa Immunoglobulin Free Light Chain as a Diagnostic Marker for Multiple Sclerosis Using Automated Chart Review

Abstract Background Multiple sclerosis (MS) is an immune-mediated central nervous system (CNS) demyelinating disease which is diagnosed based on CNS lesions' spatial and temporal dissemination. The 2017 revised McDonald Criteria include cerebrospinal fluid (CSF) specific oligoclonal bands (OCBs) to indicate dissemination in time, but their assessment is labor-intensive and requires a paired serum sample. Kappa immunoglobulin free light chain in CSF (KCSF) has emerged as a more sensitive but less specific MS biomarker, measured from a single CSF sample. Our institution's MS testing panel uses KCSF initially to rule out MS, with reflex to OCB for positive confirmation if KCSF is borderline positive (≥0.06 mg/dL). However, manual correlation of laboratory and clinical data hinders the direct comparison of the diagnostic performance between these markers. Herein we report a retrospective assessment of KCSF and OCB diagnostic performance using automated chart review. Methods Between January 1st, 2021 and December 31st, 2022, there were 1,237 patients with simultaneously measured OCB and KCSF. CSF-specific OCBs were measured by separating CSF and serum proteins using an agarose isoelectric focusing gel with a pH gradient from 3-10 (Sebia). After transfer to nitrocellulose paper, antisera were applied to visualize IgG specific bands and quantify CSF-specific OCBs. KCSF was quantitated using nephelometry by applying free light chain antisera (Binding Site) to CSF samples using a Siemens BNII analyzer. An automated chart review process utilizing keyword matching was implemented to identify patients with MS. An initial filter phase required presence of the term “multiple sclerosis”, while a second required the presence of any of the following terms: “primary progressive,” “secondary progressive,” “relapsing remitting,” “progressive relapsing,” or “tumefactive” to identify patients with diagnosed MS. The automated method’s performance was then compared to a manual chart review of 200 randomly selected patients. Subsequently, the automated process was used to determine the MS diagnosis status of all 1,237 patients. Results Of the 200 manually reviewed patients, 44 MS diagnoses were found. The automated process correctly identified 41 of these with a resulting sensitivity and specificity of 93.18% and 95.51%, respectively. Applied to the entire cohort, the automated review identified 275 MS diagnoses. For the entire cohort, the area under the curve (AUC) for the receiver operating characteristic (ROC) curves for OCB and KCSF were 0.824 (95% confidence interval: 0.794, 0.854) and 0.808 (0.777, 0.839) respectively. Our institution’s diagnostic cutoff of ≥2 for OCB yielded a sensitivity and specificity of 74.2% (68.3%, 79.1%) and 84.9% (73.8%, 89.5%), respectively. The best combination for sensitivity and specificity for KCSF with the Youden index yielded a cutoff ≥0.0654 mg/dL (sensitivity of 79.6% [73.1%, 84.0%] and specificity of 72.4% [62.8%, 76.8%]), which is similar to previous publications. This supports the use of automated chart review and suggests the KCSF cutoff has been stable over time. Conclusions KCSF exhibits comparable MS diagnostic performance to OCB yet only requires one CSF sample and can be measured rapidly. When validated, mining unstructured data allows for test performance characterization in large patient cohorts without the need for manual chart review.

Understanding privacy concerns in ChatGPT: A data-driven approach with LDA topic modeling

This study investigates privacy concerns associated with ChatGPT, a prominent generative AI model, through a data-driven approach combining Twitter data analysis and a user survey. Leveraging Latent Dirichlet Allocation (LDA) topic modeling and data categorization techniques, the research identifies key areas of concern: 1) Privacy Leakage Due to Public Data Exploitation, 2) Privacy Leakage Due to Personal Input Exploitation, and 3) Privacy Leakage Due to Unauthorized Access. Twitter data analysis of over 500k tweets, supplemented by a survey of 67 ChatGPT users, reveals nuanced user perceptions and experiences regarding privacy risks. A Python program was used to improve a dataset of 500k tweets referencing “ChatGPT” during the data preparation stage. To get a refined collection of terms, steps included converting text to lowercase, eliminating mentions and hyperlinks, tokenizing, eliminating stopwords, and keyword matching to extract tweets about ChatGPT's privacy features. Once preprocessing was completed, there were 11k refined tweets. Results highlight significant apprehensions, particularly regarding unauthorized access, underscoring the importance of robust privacy measures in AI systems. The study contributes to understanding user concerns, informing policy decisions, and guiding future research on privacy in generative AI. These studies might improve ChatGPT and other AI systems' security and privacy. The public, corporations, researchers, lawmakers, and AI developers may all benefit from the useful information it provides in better understanding and managing privacy threats.

Utilization chatbot for Indonesian tourism: a post-pandemic solution of information accessibility

Indonesia is a renowned tourist destination worldwide. However, the COVID-19 pandemic has resulted in a decrease in both foreign and domestic tourist visits. Therefore, this research aims to develop a Telegram-based ChatBot application to increase post-pandemic tourist visits. The ChatBot is designed to make it easier for tourists to obtain information related to transportation and accommodation at tourist attractions. The ChatBot is built using the Knuth-Morris-Pratt (KMP) algorithm and web scraping method. The ChatBot's response to the keyword matching is then presented to tourists. To measure travelers' perceptions of ChatBot, the System Usability Scale (SUS) was used. SUS is a questionnaire consisting of 10 questions that were answered by 20 anonymous users. Based on the calculations, the average SUS score is 71, which indicates that the developed ChatBot is in the good category and suitable for use. Using this ChatBot will make it easier for tourists to obtain tourism information in Indonesia. With its ability to retrieve relevant information related to transportation and accommodation at tourist attractions, the ChatBot can serve as a useful tool to increase post-pandemic tourist visits.

BacSPaD: A Robust Bacterial Strains' Pathogenicity Resource Based on Integrated and Curated Genomic Metadata.

The vast array of omics data in microbiology presents significant opportunities for studying bacterial pathogenesis and creating computational tools for predicting pathogenic potential. However, the field lacks a comprehensive, curated resource that catalogs bacterial strains and their ability to cause human infections. Current methods for identifying pathogenicity determinants often introduce biases and miss critical aspects of bacterial pathogenesis. In response to this gap, we introduce BacSPaD (Bacterial Strains' Pathogenicity Database), a thoroughly curated database focusing on pathogenicity annotations for a wide range of high-quality, complete bacterial genomes. Our rule-based annotation workflow combines metadata from trusted sources with automated keyword matching, extensive manual curation, and detailed literature review. Our analysis classified 5502 genomes as pathogenic to humans (HP) and 490 as non-pathogenic to humans (NHP), encompassing 532 species, 193 genera, and 96 families. Statistical analysis demonstrated a significant but moderate correlation between virulence factors and HP classification, highlighting the complexity of bacterial pathogenicity and the need for ongoing research. This resource is poised to enhance our understanding of bacterial pathogenicity mechanisms and aid in the development of predictive models. To improve accessibility and provide key visualization statistics, we developed a user-friendly web interface.

Unmasking the Contributing Factors to Oxygen Disruption Events in the Inpatient Environment and Emergency Department

Background Medical oxygen is frequently used in healthcare settings. Challenges with oxygen disruption, such as oxygen tanks running out due to communication issues between staff or tanks not being set up properly, have been noted in the limited existing literature. Challenges and patient safety issues associated with oxygen disruption persist. Utilizing a human factors approach, our study aims to understand the contributing factors and context of oxygen disruption–related patient safety event reports in the inpatient setting and provide person-based and system-based solutions. Methods Through keyword matching, we identified and then qualitatively analyzed 298 patient safety event reports to understand the factors contributing to oxygen disruption, patient location when the oxygen disruptions occurred, hand-off breakdowns by healthcare team member role, and whether high supplemental oxygen was being administered. Results The most frequent contributing factor to oxygen disruption was the patient not being transferred to another source of oxygen (n=135 of 298, 45.3%), followed by tank found empty (n=107, 35.9%), patient connected to a functioning oxygen source, no oxygen flowing (n=25, 8.4%), oxygen delivery device malfunction (n=22, 7.4%), and no oxygen available (n=9, 3.0%). Over one-third of all oxygen disruption events occurred on the unit where the patient was admitted (n=109 of 298, 36.6%). Roughly 40% of reports involved a hand-off breakdown (n=123 of 298, 41.3%) and the most frequent breakdowns occurred between a nurse and a patient transporter (n=47 of 123, 38.2%). Almost one quarter of reports involved a patient with high supplemental oxygen requirements (n=74 of 298, 24.8%). Conclusion Oxygen disruption events can have serious patient safety implications. Many of the oxygen disruption events we reviewed occurred due to lack of situational awareness and hand-off breakdowns. Combining person-based solutions, such as paper-based tools and checklists, with system-based solutions involving central monitoring and supervisory systems may help reduce the risk of oxygen disruption events.

ResumeCraft: A Machine Learning-powered Web Platform for Resume Building

Abstract: The competitive job market necessitates well-crafted resumes that resonate with both human recruiters and Applicant Tracking Systems (ATS). This paper introduces ResumeCraft, a web-based platform empowering users to build strong resumes and optimize them for ATS compatibility. ResumeCraft leverages Machine Learning (ML) for data analysis and user guidance, while the user interface is built with Hypertext Markup Language (HTML), Cascading Style Sheets (CSS), and JavaScript for a user-friendly experience. The system allows users to input their personal and professional details through a series of form fields, and provides a real-time preview of the resume design as the user inputs their data. The resume generator uses JavaScript to dynamically populate the preview with the user's input, and allows users to select from a range of pre-designed templates and color schemes to customize the look and feel of their resume. It processes the user input and generates a downloadable Portable Document Format (PDF) of the final resume. The platform analyzes user-provided information through ML models, offering suggestions on skill extraction, keyword matching, and action verb usage. This combination empowers users to create impactful resumes that are more likely to pass through ATS filters and reach human reviewers.

Cyber Bullying Detection Using Deep Learning

Abstract: Cyber bullying detection leveraging deep learning techniques. By harnessing the power of deep neural networks, specifically convolutional neural networks (CNNs) and recurrent neural networks (RNNs), we aim to develop a robust and efficient model capable of accurately identifying instances of cyberbullying in textual and multimedia content. Through extensive experimentation on diverse datasets, we demonstrate the effectiveness of our proposed method in detecting subtle forms of online harassment with high precision and recall. This paper presents an approach for cyber bullying detection through keyword analysis. With the proliferation of online platforms, identifying instances of cyberbullying has become a pressing concern. Our method leverages a predefined set of keywords associated with bullying behavior to flag potentially harmful content. Through a combination of keyword matching and contextual analysis, we demonstrate the efficacy of our approachin accurately detecting cyberbullying instances across various digital communication channels. This keyword-based detection system offers a simple yet effective means of identifying and addressing cyberbullying.

Detection of Malicious URLs using Machine Learning based on Lexical Features

As the digital world evolves, the risk of valuable information being exposed to unauthorized parties is increasing. One common vulnerability is the use of malicious Uniform Resource Locator (URL), which are fraudulent links spread across various platforms such as social media and emails. Traditional methods of identifying these URLs, such as blacklisting and heuristic search, rely heavily on syntax or keyword matching, but struggle to keep up with the evolving tactics of cyber attackers. Hence, this paper proposes a solution for detecting malicious URLs and their types based on lexical features. Lexical features in a URL refer to the components that convey semantic and lexical meaning. These can include domain names, path lengths, special characters, and other elements that can be analyzed for patterns or anomalies. In our proposed method, we use 23 different lexical features that focus on the semantic and lexical meaning of the URLs. An Exploratory Data Analysis (EDA) is used to filter the most important lexical features that effectively contribute to predictions. With these carefully curated features, we address the problem as a multi-classification task, aiming to assess the performance of three distinct classifiers: Random Forest, which currently stands as the domain's best solution and a pure bagging technique, as well as XG Boost and Light GBM, both of which utilize boosting techniques. With the proposed method, we could achieve over 93\% accuracy for all the three classifiers while Random Forest achieving the best performance.

Resume Analyzer

The aim of this project is to design and develop a tool that results into an easy and helpful solution for applicants as well as recruiters “AI RESUME ANALYZER” which parses information from a resume using natural language processing, finds the keywords, cluster them onto sectors based on their keywords And lastly show the recommendation, prediction, analytics to the applicant / recruiter based on keyword matching.

FacetDRS: Multi-faceted Feature Matching for Personalized Product Recommendations in Dynamic E-Commerce

Dynamic Product Recommender systems play a crucial role in e-commerce, guiding users towards products that align with their preferences and enhancing the shopping experience. However, existing systems often struggle to capture the nuances of user preferences, particularly when dealing with multiple features via keyword queries and filters. Additionally, scalability challenges arise as the volume of products grows with features and values. In addition, the static filters of existing recommender systems limit the exploration by strictly adhering to user selections, hindering discovery of potentially better features within the preferred budget or category. This can lead to user dissatisfaction and missed opportunities to discover more desirable options. This research addresses these limitations by introducing FacetDRS, a novel dynamic product recommendation system that leverages flexible feature matching and a scalable architecture. FacetDRS empowers the recommendations with three distinct matching methods to cater to diverse preference nuances. Exact Match ensures precise adherence to user specifications, while High-Impact Partial Match prioritizes core features with flexibility in non-critical aspects. Extended Partial Match suggests relevant options beyond initial preferences, leveraging predefined thresholds. This multi-faceted approach captures the full spectrum of user desires while maintaining accuracy. To evaluate the effectiveness of FacetDRS, a rigorous experimental setup was conducted, comparing its performance across Keyword Match Score, First Selections in Top-K and Average Response Time metrics with established recommender systems such as DeepRec, APGNN, and TADCF. The results demonstrate that FacetDRS achieves superior performance than its counterparts in all comparison aspects. The system's flexible matching methods, combined with its robust architecture, offer a valuable solution for enhancing user satisfaction and driving conversions in the mobile commerce landscape.

Manufacturing service capability prediction with Graph Neural Networks

In the current landscape, the predominant methods for identifying manufacturing capabilities from manufacturers rely heavily on keyword matching and semantic matching. However, these methods often fall short by either overlooking valuable hidden information or misinterpreting critical data. Consequently, such approaches result in an incomplete identification of manufacturers’ capabilities. This underscores the pressing need for data-driven solutions to enhance the accuracy and completeness of manufacturing capability identification. To address the need, this study proposes a Graph Neural Network-based method for manufacturing service capability identification over a knowledge graph. To enhance the identification performance, this work introduces a novel approach that involves aggregating information from the graph nodes’ neighborhoods as well as oversampling the graph data, which can be effectively applied across a wide range of practical scenarios. Evaluations conducted on a Manufacturing Service Knowledge Graph and subsequent ablation studies demonstrate the efficacy and robustness of the proposed approach. This study not only contributes a innovative method for inferring manufacturing service capabilities but also significantly augments the quality of Manufacturing Service Knowledge Graphs.

Resume Optimisation and Suggestions using Large Language Models: A Review

Abstract: In the rapidly evolving job market, the compatibility of a candidate's resume with the job description is critical to the application process. Conventional resume optimization techniques frequently depend on keyword matching, which is shallow and cannot comprehend linguistic nuances or context, resulting in an inadequate match between applicant's profiles and job criteria. This study presents a novel application that uses Bidirectional Encoder Representations from Transformers (BERT) to generate language embeddings from job descriptions and resumes, therefore bridging this gap. Our program offers a contextaware and nuanced evaluation of how well a candidate's qualifications match the requirements of the position by calculating the similarity scores between these embeddings. Furthermore, the incorporation of Generative Pre-trained Transformer (GPT) models provides customized suggestions for resume optimization, emphasizing areas of improvement indicated by the BERT study. Additionally, the application also has a job search tool that lets the users browse LinkedIn for suitable job posts and apply straight to them.

The antimicrobial effect of Limosilactobacillus reuteri as probiotic on oral bacteria: A scoping review

Dysbiosis among oral microbial community in the oral cavity can lead to several oral diseases. Probiotic therapy is known to correct these imbalances. Limosilactobacillus reuteri is one of the most studied strains of probiotics and can control oral microbiota through reuterin, a wide-spectrum antimicrobial agent. The objective of this review was to evaluate the effect of the antimicrobial activity of Limosilactobacillus reuteri on the oral bacteria of humans. This review used PubMed, Scopus, EMBASE, ScienceDirect, and Google Scholar databases as bibliographic resources. Studies with matching keywords were analyzed and screened with PRISMA-ScR recommendations. Sixteen articles were selected for this review, which included a total of 832 patients. Based on this review, Limosilactobacillus reuteri has a strong antibacterial effect against Streptococcus mutans in healthy individuals but is not effective against Lactobacillus. Additionally, it has a significant antibacterial effect against Porphiromonas gingivalis in patients with periodontitis, although its effectiveness is not stable in patients with peri-implant infections. Furthermore, Limosilactobacillus reuterihas varying results against other bacteria, indicating the need for further extensive research to ensure its efficacy.

Development and Evaluation of Surveillance System for Identifying Jail-Associated COVID-19 Cases in Minnesota, USA, 2022.

Confinement facilities are high-risk settings for the spread of infectious disease, necessitating timely surveillance to inform public health action. To identify jail-associated COVID-19 cases from electronic laboratory reports maintained in the Minnesota Electronic Disease Surveillance System (MEDSS), Minnesota, USA, the Minnesota Department of Health developed a surveillance system that used keyword and address matching (KAM). The KAM system used a SAS program (SAS Institute Inc., https://www.sas.com) and an automated program within MEDSS to identify confinement keywords and addresses. To evaluate KAM, we matched jail booking data from the Minnesota Statewide Supervision System by full name and birthdate to the MEDSS records of adults with COVID-19 for 2022. The KAM system identified 2,212 cases in persons detained in jail; sensitivity was 92.40% and specificity was 99.95%. The success of KAM demonstrates its potential to be applied to other diseases and congregate-living settings for real-time surveillance without added reporting burden.

The application of NLP in information retrieval

The field of Natural Language Processing (NLP) has experienced impressive advancements and has found diverse applications. This paper presents a comprehensive review of the development of NLP in the field of information retrieval. It explores different stages of NLP techniques and methods, including keyword matching, rule-based approaches, statistical methods, and the utilization of machine learning and deep learning technologies. Furthermore, the paper provides detailed insights into the specific applications of NLP in domains such as academic information retrieval, medical information retrieval, travel information retrieval, and e-commerce information retrieval. It analyzes the current state of NLP applications in these domains, highlights their advantages, and discusses their associated limitations. Finally, the paper emphasizes the continuous advancement of the NLP field, with a particular focus on semantic understanding, personalized retrieval, and multimodal information retrieval, to better adapt to diverse data types and user requirements. The paper concludes by summarizing the main points discussed and providing future directions.

Intelligent material classification of traditional Chinese medicine based on semantic analysis

A method for material classification of traditional Chinese medicines based on the physical properties of powder has been established by our research group. This method involves pre-treatment of traditional Chinese medicine decoction pieces, powder preparation, and determination of physical properties, being cumbersome. In this study, the word segmentation logic of semantic analysis was adopted to establish the thesaurus and local standardized semantic word segmentation database with the macroscopic and microscopic characteristics of 36 model traditional Chinese medicines as the basic data. The physical properties of these medicines have been determined and the classification of these medicines is clear in the cluster analysis. A total of 55 keywords for powdery, fibrous, sugary, oily, and brittle materials were screened by association rules and the set inclusion and exclusion criteria, and the weights of the keywords were calculated. Furthermore, the algorithms of the keyword matching scores and the computation rules of the single or multiple material classification were established for building the intelligent model of semantic analysis for the material classification. The semantic classification results of the other 35 TCMs except Pseudostellariae Radix(multi-material medicine) agreed with the clustering results based on the physical properties of the powder, with an agreement rate of 97.22%. In model validation, the prediction results of semantic classification of traditional Chinese medicines were consistent with the clustering results based on the physical properties of powder, with an agreement rate of 83.33%. The results showed that the method of material classification based on semantic analysis was feasible, which laid a foundation for the development of intelligent decision-making technology for personalized traditional Chinese medicine preparations.

A systematic review of efficacy on larva debridement in diabetic foot ulcers

Larva therapy, also known as maggot therapy, biodebridement, or maggot debridement therapy, involved the application of live fly larva to a patient’s wound for debridement, disinfection, and wound healing. Despite its wide application in diabetic foot ulcers, the efficacy of this intervention remains uncertain. This review aimed to examine the benefits of larva therapy in diabetic foot ulcers by conducting systematic review. This systematic review sought relevant articles using MeSH-based keywords in databases such as Scopus, PubMed, Science Direct, and ProQuest. The inclusion criteria for the articles were as follows: population - patients with diabetic foot ulcers, intervention - the use of debridement or therapeutic application of maggot therapy with Lucilia sericata, comparison - none, outcome - the results of articles demonstrating the effectiveness of using Lucilia sericata maggot therapy on diabetic foot ulcers. Article quality assessment was conducted using the Joanna Briggs Institute guidelines. Descriptive analysis was performed with a narrative approach, considering articles published from 2018 to 2023. Based on the literature search, 237 articles were found with matching keywords from Scopus (103), PubMed (77), Science Direct (14), and ProQuest (43). After screening for inclusion and exclusion criteria, 11 articles were identified for inclusion in this review. These 11 articles indicate that larva therapy can be an effective method in treating diabetic foot ulcers, especially when used correctly and tailored to each patient’s condition and needs. Green bottle fly larva, Lucilia sericata, produce proteolytic enzymes that can assist in wound healing. Larva therapy is effective when combined with other procedures. In clinical practice, it is highly recommended that nurses provide an individualized approach and assess potential side effects.

Construction and Innovative Application of Fundamental Training Educational Resource Repository Using Data Mining Techniques

With the rapid advancement of information technology and data mining methodologies, the educational sector is undergoing an unprecedented transformation. The increasing proliferation of online educational resources offers significant advantages for educators and learners, yet presents challenges in resource retrieval and recommendation. Existing query and recommendation systems frequently rely on simplistic keyword matches and browsing history, struggling to cater to the escalating demand for personalized learning experiences. This study introduces an innovative method for constructing a Fundamental Training Educational Resource Repository (FTERR) leveraging data mining techniques. Firstly, students’ learning needs are intricately captured using knowledge graph and semantic analysis methods. Secondly, addressing the challenges of resource storage and retrieval, a novel data storage model is proposed, which, when paired with cutting-edge clustering algorithms, facilitates rapid and visualized retrieval of educational materials. This research not only furnishes novel methodologies for data mining applications in the educational domain but also fosters the advancement of educational informatization.

OEQA: Knowledge- and Intention-Driven Intelligent Ocean Engineering Question-Answering Framework

The constantly updating big data in the ocean engineering domain has challenged the traditional manner of manually extracting knowledge, thereby underscoring the current absence of a knowledge graph framework in such a special field. This paper proposes a knowledge graph framework to fill the gap in the knowledge management application of the ocean engineering field. Subsequently, we propose an intelligent question-answering framework named OEQA based on an ocean engineering-oriented knowledge graph. Firstly, we define the ontology of ocean engineering and adopt a top-down approach to construct a knowledge graph. Secondly, we collect and analyze the data from databases, websites, and textual reports. Based on these collected data, we implement named entity recognition on the unstructured data and extract corresponding relations between entities. Thirdly, we propose an intent-recognizing-based user question classification method, and according to the classification result, construct and fill corresponding query templates by keyword matching. Finally, we use T5-Pegasus to generate natural answers based on the answer entities queried from the knowledge graph. Experimental results show that the accuracy in finding answers is 89.6%. OEQA achieves in the natural answer generation in the ocean engineering domain significant improvements in relevance (1.0912%), accuracy (4.2817%), and practicability (3.1071%) in comparison to ChatGPT.