Keyword-based Approach Research Articles

AI-Powered Search Systems : Integrating Machine Learning with Search Technology for High-Scalability Applications

This article comprehensively analyzes integrating Artificial Intelligence (AI) and machine learning techniques into high-scalability search systems. We explore AI-powered search's theoretical foundations and practical implementations, focusing on advanced ranking algorithms, natural language processing for query understanding, and optimized distributed architectures. We demonstrate significant improvements in search relevance and efficiency through experiments conducted on a large-scale dataset comprising 100 million web pages and 1 million real-world queries. Our AI-powered system showed a 15% increase in Normalized Discounted Cumulative Gain (NDCG) for complex queries and a 12% improvement in Mean Reciprocal Rank (MRR) for navigational queries compared to traditional keyword-based approaches. We also address critical challenges in maintaining system scalability and performance, including data synchronization, real-time model updates, and resource management in distributed environments. The article further discusses emerging trends, such as graph neural networks and multimodal search capabilities, alongside ethical considerations and data privacy concerns. Our findings provide valuable insights for researchers and practitioners aiming to develop next-generation search platforms capable of handling the increasing complexity and volume of digital information while ensuring responsible AI integration.

Unraveling the Nuclear Debate: Insights Through Clustering of Tweets

The perception of nuclear power, while central to energy policy and sustainability endeavors, remains a subject of considerable debate, in which some claim that the expansion of nuclear technology poses threats to global security, while others argue that its access should be shared for development and energy purposes. In this study, a total of 11,256 tweets were gathered over a three-month period using a keyword-based approach through the Twitter Standard Search API, focusing on terms related to nuclear energy. The k-means clustering algorithm was employed to analyze tweets with the aim of determining the underlying sentiments and perspectives within the public domain, while t-SNE was used for visualizing cluster separation. The results show distinct clusters reflecting various viewpoints on nuclear power, with 71.94% of tweets being neutral, 14.64% supportive, and 13.42% negative. This study also identifies a subset of users who appear to be seeking unbiased information, signaling an opportunity for educational outreach. By leveraging the immediacy and pervasiveness of X (formerly known as Twitter), this research provides a timely snapshot of the prevailing attitudes toward nuclear power and offers insights for policymakers, educators, and industry stakeholders.

Measuring Cybercrime in Calls for Police Service

Conventional police databases contain much information on cybercrime, but extracting it remains a practical challenge. This is because these databases rarely contain labels that could be used to automatically retrieve all cybercrime incidents. In this article, we present a supervised machine learning method for extracting cybercrime incidents in calls for police service datasets. Data from the Korean National Police (2020, 9 months, N = 15 million call logs) is used for the demonstration. We combined methods of keyword query selection, minority oversampling, and majority voting techniques to develop a classifier. Three classification techniques, including Naïve Bayes, linear SVM, and kernel SVM, were tested, and the kernel model was chosen to build the final model (accuracy, 93.4%; F1-score, 92.4). We estimate that cybercrime only represents 4.6% of the cases in the selected dataset (excluding traffic-related incidents), but that it can be prevalent with some crime types. We found, for example, that about three quarters (76%) of all fraud incidents have a cyber dimension. We conclude that the cybercrime classification method proposed in this study can support further research on cybercrime and that it offers considerable advantages over manual or keyword-based approaches.

Towards Effective and Efficient Sparse Neural Information Retrieval

Sparse representation learning based on Pre-trained Language Models has seen a growing interest in Information Retrieval. Such approaches can take advantage of the proven efficiency of inverted indexes and inherit desirable IR priors such as explicit lexical matching or some degree of interpretability. In this work, we thoroughly develop the framework of sparse representation learning in IR, which unifies term weighting and expansion in a supervised setting. We then build on SPLADE—a sparse expansion-based retriever—and show to which extent it is able to benefit from the same training improvements as dense bi-encoders by studying the effect of distillation, hard negative mining, as well as the Pre-trained Language Model’s initialization on its effectiveness , leading to state-of-the-art results in both in- and out-of-domain evaluation settings (SPLADE++). We furthermore propose efficiency improvements, allowing us to reach latency requirements on par with traditional keyword-based approaches (Efficient-SPLADE).

Repairing raw metadata for metadata management

With the exponential growth of data production, the generation of metadata has become an integral part of the process. Metadata plays a crucial role in facilitating enhanced data analytics, data integration, and resource management by offering valuable insights. However, inconsistencies arise due to deviations from standards in metadata recording, including missing attribute information, publishing URLs, and provenance. Furthermore, the recorded metadata may exhibit inconsistencies, such as varied value formats, special characters, and inaccurately entered values. Addressing these inconsistencies through metadata preparation can greatly enhance the user experience during data management tasks.This paper introduces MDPrep, a system that explores the usability and applicability of data preparation techniques in improving metadata quality. Our approach involves three steps: (1) detecting and identifying problematic metadata elements and structural issues, (2) employing a keyword-based approach to enhance metadata elements and a syntax-based approach to rectify structural metadata issues, and (3) comparing the outcomes to ensure improved readability and reusability of prepared metadata files.

An Experiment in Keyword-based Approach for Hate Speech Detection in Online Social Media Comments

An Experiment in Keyword-based Approach for Hate Speech Detection in Online Social Media Comments

Blockchain in the Energy Sector for SDG Achievement

Blockchain technology finds application in multiple sectors, including renewable energy. Numerous blockchain-based applications aim to provide support in the production, management, distribution, and consumption of green energy. The benefits offered are not only technological but also social, environmental, and economic. The purpose of this study is to examine how the application of blockchain in the energy industry may affect the achievement of the Sustainable Development Goals (SDGs). This study is composed of two parts. The first part concerns the identification and analysis of the most relevant categories of blockchain applications in the energy sector and their ability to contribute to the achievement of the SDGs. A knowledge base, comprising scientific articles, gray literature, and real-world applications, has been created and analyzed. With a keyword-based approach, each application was associated with one or more SDGs. In the second part, the Sustainability Awareness Framework (SuSAF) was used to examine the findings of the first part of the study and discuss them in terms of five dimensions of sustainability. Finally, potential risks associated with the use of blockchain in the energy sector are also covered. Results reveal that tracking energy production and consumption and renewable energy communities are the applications that have the most beneficial effects, and that the benefits linked to blockchain adoption go beyond the energy sector to include the environment, the economy, industry, infrastructure, smart cities, and society.

Sentiment spin: Attacking financial sentiment with GPT-3

In this study, we explore the susceptibility of financial sentiment analysis to adversarial attacks that manipulate financial texts. With the rise of AI readership in the financial sector, companies are adapting their language and disclosures to fit AI processing better, leading to concerns about the potential for manipulation. In the finance literature, keyword-based methods, such as dictionaries, are still widely used for financial sentiment analysis due to their perceived transparency. However, our research demonstrates the vulnerability of keyword-based approaches by successfully generating adversarial attacks using the sophisticated transformer model, GPT-3. With a success rate of nearly 99% for negative sentences in the Financial Phrase Bank, a widely used database for financial sentiment analysis, we highlight the importance of incorporating robust methods, such as context-aware approaches such as BERT, in financial sentiment analysis.

Using social media to analyze consumers' attitude toward natural food products

PurposeThis paper examines user generated social media content bearing on consumers’ attitude and belief systems taking the domain of natural food product as illustrative case. This research sheds light on how consumers think and talk about natural food within the context of food well-being and health.Design/methodology/approachThe authors used a keyword-based approach to extract user generated content from Twitter and used both food as well-being and food as health frameworks for analysis of more than two million tweets.FindingsThe authors found that consumers mostly discuss food marketing and less frequently discuss food policy. Their results show that tweets regarding naturalness were significantly less frequent in food categories that feature naturalness to an extent, e.g. fruits and vegetables, compared to food categories dominated by technologies, processing and man-made innovation, such as proteins, seasonings and snacks.Research limitations/implicationsThis paper provides numerous implications and contributions to the literature on consumer behavior, marketing and public policy in the domain of natural food.Practical implicationsThe authors’ exploratory findings can be used to guide food system stakeholders, farmers and food processors to obtain insights into consumers' mindset on food products, novel concepts, systems and diets through social media analytics.Originality/valueThe authors’ results contribute to the literature on the use of social media in food marketing on understanding consumers' attitudes and beliefs toward natural food, food as the well-being literature and food as the health literature, by examining the way consumers think about natural (versus man-made) food using user generated content of Twitter, which has not been previously used.

Knowledge-based System of Indian Culture using Ontology with Customized Named Entity Recognition

Indian culture is a blend of diversified ethics with different cultures and religions. There arise the sustainable challenges for the people of inter societies and religions to participate in the regional events with lack of cultural information like types of cuisines eaten; rituals followed etc. The existing search engines using a keyword-based approach make it difficult for people to find relevant information among millions of undesirable results. To overcome this problem, we applied the concept of the semantic web that facilitates unstructured data that is represented more precisely by using well-defined vocabularies such as Resource Description Framework (RDF), which is recognized by both people and machines. In this research, we have developed a Knowledge Base System (KBS) using an ontological approach (Ontology Development 101) integrating the concept of Customized Named Entity Recognition (CNER) that supports full-text query searching with structured information. To test the feasibility, several queries from our domain taken from different users were executed successfully. The precision of 0.96 justifies the efficacy of our framework. Finally, the evaluation results were compared with two previous methodologies which show that our methodology outperforms.

On the documentation of refactoring types

Commit messages are the atomic level of software documentation. They provide a natural language description of the code change and its purpose. Messages are critical for software maintenance and program comprehension. Unlike documenting feature updates and bug fixes, little is known about how developers document their refactoring activities. Specifically, developers can perform multiple refactoring operations, including moving methods, extracting classes, renaming attributes, for various reasons, such as improving software quality, managing technical debt, and removing defects. Yet, there is no systematic study that analyzes the extent to which the documentation of refactoring accurately describes the refactoring operations performed at the source code level. Therefore, this paper challenges the ability of refactoring documentation, written in commit messages, to adequately predict the refactoring types, performed at the commit level. Our analysis relies on the text mining of commit messages to extract the corresponding features (i.e., keywords) that better represent each class (i.e., refactoring type). The extraction of text patterns, specific to each refactoring type (e.g., rename, extract, move, inline, etc.) allows the design of a model that verifies the consistency of these patterns with their corresponding refactoring. Such verification process can be achieved via automatically predicting, for a given commit, the method-level type of refactoring being applied, namely Extract Method, Inline Method, Move Method, Pull-up Method, Push-down Method, and Rename Method. We compared various classifiers, and a baseline keyword-based approach, in terms of their prediction performance, using a dataset of 5004 commits. Our main findings show that the complexity of refactoring type prediction varies from one type to another. Rename Method and Extract Method were found to be the best documented refactoring activities, while Pull-up Method, and Push-down Method were the hardest to be identified via textual descriptions. Such findings bring the attention of developers to the necessity of paying more attention to the documentation of these types.

Using Ontology-Based Approach to Improved Information Retrieval Semantically for Historical Domain

Searching and retrieving documents from large historical archives prove to be challenging for the information retrieval (IR) field as historians typically employ their knowledge, experience, and intuition. There are several works done on the application of IR in historical documents. As such, the conventional IR model is mostly used a simple Bag-of-Word (BOW) approach and usually unable to support precise document retrieval for the domain of history. We proposed an ontology-based approach to semantically index and ranked rich historical documents. The historical documents relating to the Vietnam War were chosen for this study. Several existing ontologies have been reviewed to identify the most suitable concepts and properties which contain rich information pertaining to relevant entities such as an event, time, and people. The domain ontology was developed by utilizing the existing Simple News and Press (SNaP) ontology and extended with concepts related to the Vietnam War. The ontology was then semantically mapped with concepts found in a collection of 133 documents relating to the Vietnam war. In this paper, we also proposed a simple ontology-based weighting mechanism derived from the classic tf-idf scoring scheme. Finally, 20 SPARQL queries are implemented to do the evaluation. The evaluation shows that the proposed ontological-based approach achieved better results as compared to the base-line BM-25 probabilistic retrieval model in terms of precision and recall metrics. The use of the ontology-based approach in document retrieval can compete with the keyword-based approach.

Bridging Qualitative and Quantitative Methods for User Modeling: Tracing Cancer Patient Behavior in an Online Health Community

Researchers construct models of social media users to understand human behavior and deliver improved digital services. Such models use conceptual categories arranged in a taxonomy to classify unstructured user text data. In many contexts, useful taxonomies can be defined via the incorporation of qualitative findings, a mixed-methods approach that offers the ability to create qualitatively-informed user models. But operationalizing taxonomies from the themes described in qualitative work is non-trivial and has received little explicit focus. We propose a process and explore challenges bridging qualitative themes to user models, for both operationalization of themes to taxonomies and the use of these taxonomies in constructing classification models. For classification of new data, we compare common keyword-based approaches to machine learning models. We demonstrate our process through an example in the health domain, constructing two user models tracing cancer patient experience over time in an online health community. We identify patterns in the model outputs for describing the longitudinal experience of cancer patients and reflect on the use of this process in future research.

LTRWES: A new framework for security bug report detection

Abstract Context: Security bug reports (SBRs) usually contain security-related vulnerabilities in software products, which could be exploited by malicious attackers. Hence, it is important to identify SBRs quickly and accurately among bug reports (BRs) that have been disclosed in bug tracking systems. Although a few methods have been already proposed for the detection of SBRs, challenging issues still remain due to noisy samples, class imbalance and data scarcity. Object: This motivates us to reveal the potential challenges faced by the state-of-the-art SBRs prediction methods from the viewpoint of data filtering and representation. Furthermore, the purpose of this paper is also to provide a general framework and new solutions to solve these problems. Method: In this study, we propose a novel approach LTRWES that incorporates learning to rank and word embedding into the identification of SBRs. Unlike previous keyword-based approaches, LTRWES is a content-based data filtering and representation framework that has several desirable properties not shared in other methods. Firstly, it exploits ranking model to efficiently filter non-security bug reports (NSBRs) that have higher content similarity with respect to SBRs. Secondly, it applies word embedding technology to transform the rest of NSBRs, together with SBRs, into low-dimensional real-value vectors. Result: Experiment results on benchmark and large real-world datasets show that our proposed method outperforms the state-of-the-art method. Conclusion: Overall, the LTRWES is valid with high performance. It will help security engineers to identify SBRs from thousands of NSBRs more accurately than existing algorithms. Therefore, this will positively encourage the research and development of the content-based methods for security bug report detection.

Identifying security issues for mobile applications based on user review summarization

Abstract Context With the development of mobile apps, public concerns about security issues are continually rising. From the user’s perspective, it is crucial to be aware of the security issues of apps. Reviews serve as an important channel for users to discover the diverse issues of apps. However, previous works rarely rely on existing reviews to provide a detailed summarization of the app’s security issues. Objective To provide a detailed overview of apps’ security issues for users, this paper introduces SRR-Miner, a novel review summarization approach that automatically summarizes security issues and users’ sentiments. Method SRR-Miner follows a keyword-based approach to extracting security-related review sentences. It summarizes security issues and users’ sentiments with triples, which makes full use of the deep analysis of sentence structures. SRR-Miner also provides visualized review summarization through a radar chart. Results The evaluation on 17 mobile apps shows that SRR-Miner achieves higher F1-score and MCC than Machine Learning-based classification approaches in extracting security-related review sentences. It also accurately identifies misbehaviors, aspects and opinions from review sentences. A qualitative study shows that SRR-Miner outperforms two state-of-the-art approaches (AR-Miner and SUR-Miner) in terms of summarizing security issues and users’ sentiments. A further user survey indicates the usefulness of the summarization of SRR-Miner. Conclusion SRR-Miner is capable of automatically extracting security-related review sentences based on keywords, and summarizing misbehaviors, aspects and opinions of review sentences with a deep analysis of the sentence structures.

Language and Obfuscation Oblivious Source Code Authorship Attribution

Source Code Authorship Attribution can answer many interesting questions such as: Who wrote the malicious source code? Is the source code plagiarized, and does it infringe on copyright? Source Code Authorship Attribution is done by observing distinctive patterns of style in a source code whose author is unknown and comparing them with patterns learned from known authors' source codes. In this paper, we present an efficient approach to learn a novel representation using deep metric learning. The existing state of the art approaches tokenize the source code and work on the keyword level, limiting the elements of style they can consider. Our approach uses the raw character stream of source code. It can examine keywords and different stylistic features such as variable naming conventions or using tabs vs. spaces, enabling us to learn a richer representation than other keyword-based approaches. Our approach uses a character-level Convolutional Neural Network (CNN). We train the CNN to map the input character stream to a dense vector, mapping the source codes authored by the same author close to each other. In contrast, source codes written by different programmers are mapped farther apart in the embedding space. We then feed these source code vectors into the K-nearest neighbor (KNN) classifier that uses Manhattan-distance to perform authorship attribution. We validated our approach on Google Code Jam (GCJ) dataset across three different programming languages. We prepare our large-scale dataset in such a way that it does not induce type-I error. Our approach is more scalable and efficient than existing methods. We were able to achieve an accuracy of 84.94% across 20,458 authors, which is more than twice the scale of any previous study under a much more challenging setting.

Keyword-based approach for recognizing fraudulent messages by keystroke dynamics

In recent years, many approaches that use keystroke dynamics in free text authentication have been proposed. The major drawback of the proposed approaches is that training generally requires several months, thereby resulting in low practicality. In this study, a method to detect U.S. English fraudulent messages by analyzing keyboard users' keystroke dynamics is proposed. To the best of our knowledge, this is the first study to apply keystroke dynamics to detect fraudulent instant messages. In the proposed system, each user requires only approximately 20 min of training in U.S. English keystroke dynamics. Furthermore, a voting-based statistical classifier is presented to improve the recognition accuracy of instant messages and prevent phishing messages. Experimental results indicate that the proposed approach outperforms other relevant published methods in terms of shorter training time, fewer false alarms, and comparable recognition accuracy.

A positional keyword-based approach to inferring fine-grained message formats

Message format extraction, the process of revealing the message syntax without access to the protocol specification, is important for a variety of applications such as service virtualization and network security. In this paper, we propose P-token, which mines fine-grained message formats from network traces. The novelty of our approach is twofold: a ‘positional keyword’ identification technique and a two-level hierarchical clustering strategy. Positional keywords are based on the insight that keywords or reserved words usually occur at relatively fixed positions in the messages. By associating positions as meta-information with keywords, we can more accurately distinguish keywords from message payload data. After identification, the positional keywords are used as features to cluster the messages using density peaks clustering. We then perform another level of clustering to refine the clusters with low homogeneity. Finally, the message format of each cluster is extracted based on the observed ordering of keywords. P-token improves on the current state-of-the-art techniques by successfully addressing two challenges that commonly afflict existing keyword based format extraction methods: message keyword mis-identification and message format over-generalization. We have conducted experiments on services and applications using various protocols, including SOAP, LDAP, IMS and a RESTful service. Our experimental results show that P-token outperforms existing methods in extracting message formats.

A Novel IR for Relational Database using Optimize Query Building

The different type search engine like Google, binge, AltaVista is used to fetch the information from the database by easy language. The non-technical employee they don’t understand the database and query cannot access the database. The proposed system is performing work as a search engine where users can fetch the information from the database by natural human sounding language. The previous existing system doesn’t able to solve queries in one easy statement. The structured query approach, while expressive and powerful, is not easy for naive users. The keyword-based approach is very friendly to use, but cannot express complex query intent accurately. This paper emphasis on Natural Language based query processor. We have proposed the use of query optimization approach to convert complex NLP query to SQL query, SPAM word removal, POS tagger applied over NL query and concluded that execution time lesser when query size increases.

Automatic topic labelling for text document using Ontology of graph-based concepts and dependency graph

Topic labelling is an important task of text mining. It supports assigning proper topic labels to the text documents. In this paper, we present a novel approach of using graph-based concept matching approach in solving automatic topic labelling task. Our proposed model demonstrates that the quality of automatic topic labelling task for text documents can be improved, in comparison with traditional keyword-based concept matching approach. In this paper, we propose a novel approach of automatic ontology-driven topic labelling. Our proposed model is considered as a semi-supervised approach. It uses existed ontologies as the pre-knowledge base for topic identification in text documents. We performed the experiments on the real-world ACM's documents to show the effectiveness of our proposed model in solving topic labelling task. The experimental results on real-world and standard datasets demonstrate that our proposed model can leverage the output accuracy of topic labelling task in text documents.