Personalized Information Retrieval Research Articles

Design and Implementation of an Interactive Question-Answering System with Retrieval-Augmented Generation for Personalized Databases

This study introduces a novel approach to personalized information retrieval by integrating retrieval augmentation generation (RAG) with a personalized database system. Recent advancements in large language models (LLMs) have shown impressive text generation capabilities but face limitations in knowledge accuracy and hallucinations. Our research addresses these challenges by combining LLMs with structured, personalized data to enhance search precision and relevance. By tagging keywords within personal documents and organizing information into context-based categories, users can conduct efficient searches within their data repositories. We conducted experiments using the GPT-3.5 and text-embedding-ada-002 models and evaluated the RAG assessment framework with five different language models and two embedding models. Our results indicate that the combination of GPT-3.5 and text-embedding-ada-002 is effective for a personalized database question-answering system, with potential for various language models depending on the application. Our approach offers improved accuracy, real-time data updates, and enhanced user experience, making a significant contribution to information retrieval by LLMs and impacting various artificial intelligence applications.

Application Research of Intelligent Search System Based on Clustering Algorithm

In this paper, a multi-agent oriented personalized information retrieval method is proposed. The system is composed of 5 agents and 2 other components. The design highlights the personalization and real-time of the information obtained by users, and has the characteristics of intelligent retrieval and active retrieval. Then this paper studies it from three aspects: user model, system learning model and individual information model. Then a grouping gravity search algorithm is proposed. Compared with conventional gravity search, group gravity search is carried out for a specific decoding strategy. This algorithm is a new gravity search algorithm suitable for packet coding, which makes the packet search process approximate to the iterative optimization of classical gravity algorithm. Finally, the effectiveness of the proposed method is verified on various typical experimental cases. Experiments show that the proposed algorithm has higher classification performance than other intelligent group algorithms.

Evaluation of Natural Language Processing Techniques for Information Retrieval

Purpose: The main objective of the study was to investigate the evaluation of Natural Language Processing techniques for information retrieval.
 Methodology: The study adopted a desktop research methodology. Desk research refers to secondary data or that which can be collected without fieldwork. Desk research is basically involved in collecting data from existing resources hence it is often considered a low cost technique as compared to field research, as the main cost is involved in executive’s time, telephone charges and directories. Thus, the study relied on already published studies, reports and statistics. This secondary data was easily accessed through the online journals and library.
 Findings: The findings reveal that there exists a contextual and methodological gap relating to Natural Language Processing techniques for information retrieval. Preliminary empirical review revealed that NLP methods significantly improved the accuracy and efficiency of information retrieval systems. Through systematic evaluation, various NLP techniques, including tokenization, named entity recognition, semantic parsing, and word embeddings, were found to enhance retrieval performance across diverse datasets and domains. By considering context and user intent, researchers aimed to develop more contextually aware and personalized information retrieval systems. The study emphasized the need for further research to explore hybrid approaches and domain-specific adaptations, ultimately highlighting the transformative potential of NLP in revolutionizing information access and utilization.
 Unique Contribution to Theory, Practice and Policy: Information Foraging theory, Relevance theory and Cognitive Load theory may be used to anchor future studies on Natural Language Processing techniques. The study provided recommendations to enhance information retrieval systems. It suggested integrating advanced NLP techniques such as named entity recognition and sentiment analysis to improve query understanding and document relevance. Additionally, the study recommended adopting word embeddings and semantic parsing techniques for better semantic understanding of user queries and documents. It emphasized the importance of domain-specific adaptations and continuous evaluation of NLP techniques to tailor them to specific application domains and keep pace with evolving user needs and technological advancements.

Is it the new Google: Impact of ChatGPT on Students' Information Search Habits

The traditional subject of information search and retrieval (IR) paradigm shifted to an entirely new era since artificial intelligence (AI) techniques were introduced into the field. Browser-based IR solutions powered by AI for personalised recommendations-based information retrieval, such as the Google search engine, were one of the early examples. The IR field has advanced to its next level with the newest conversational applications based on large language model (LLM) techniques. It is becoming clear that Generative Pretrained Transformer (GPT) applications such as ChatGPT will significantly impact information retrieval behaviour in the education sector. Though this application has become widespread in acclaim, no previous study has shown its impact on information seeking and retrieval. However, based on the observation of the fast penetration of this technology and the growth of public interest, a pre-assumption was built that it is essential to investigate if students may also be showing a similar interest in this new tool. Hence, this study is set up to systematically and empirically explore how ChatGPT influences the IR behaviour of students in HEIs. A survey approach is utilised to collect the perceived IR behaviour through a questionnaire administered to 60 students in HEIs. The findings reveal that the tool is already widely known among HEI students. They also perceived the use of the tool in the context of information retrieval and proclaimed its usefulness, acknowledging its efficiency (reduced time) in finding information. Furthermore, the technology has considerably affected the typical use of other conventional information retrieval and search engine tools. On the contrary, 10% of the respondents are less likely to use ChatGPT during information seeking for various reasons, from credibility and relevance to technology infrastructure issues such as connectivity. Although a deeper analysis is required to establish a general conclusion on how and in which ways GPT-based models will override contemporary IR practices, the study outcome provides evidence for a possible behavioural change among HEI students in their IR habits in the future.

Enhancing Class Management in Chinese Schools Through Semantic Web Technologies

This paper explores the potential of utilizing semantic web technologies to improve class management in Chinese schools. By analyzing a comprehensive dataset obtained from the Scopus database, the study investigates publication trends, document types, keyword distributions, and author contributions in the field of semantic web technologies for class management. The findings reveal a growing interest in this research area and highlight the benefits of semantic web technologies in personalized learning, information retrieval, collaboration, and assessment. The paper discusses the practical implications, challenges, and considerations for implementing semantic web technologies in Chinese schools. It aims to provide valuable insights for educators, researchers, policymakers, and educational technology practitioners interested in enhancing class management practices through the innovative use of semantic web technologies.

Exploring the Links of Mobility-as-a-Service Platform Functionalities

Mobility as a Service (MaaS) includes services related to public and other shared mobility means of transport via an integrated mobile platform. Despite the existing initiatives and research regarding the optimal sets of platform features, the relationships and importance of these features (when seen as a network) are still under investigation. This study explores the relationships of MaaS platform features from the perspective of mobility service providers (MSPs) and commuters. We collected data via a focus group with five mobility experts and a survey (71 Dutch commuters). Our results show that optimizing the number and use of the vehicles and appropriate data handling are the essential features of a MaaS platform for the MSPs. Although retrieval of personal information and additional services (e.g., concert tickets) are welcome, they are considered less significant for the commuters.

MaaS platform features: An exploration of their relationship and importance from supply and demand perspective

Mobility as a Service (MaaS) aims to offer travelers easy and convenient access to transportation modes via a joint digital channel, often in a mobile application. MaaS has the potential to fundamentally change the way we commute as it encourages a more sustainable travel behavior. MaaS, as a business model, is an innovative concept that integrates a variety of these solutions and can significantly change the mobility environment by encouraging sustainable travel behavior. Despite the existing initiatives and efforts toward MaaS solutions, there is still no consensus on the essential features of MaaS platforms for actors from the supply and demand sides. This study explores the critical features of MaaS platforms from the perspective of mobility service providers - MSP (i.e., supply-side) and travelers (i.e., demand-side). We collected data via interviews with mobility experts for the supply side and a survey for the users' side. Based on a Gaussian Graphical Model, our results show that optimizing the number and use of the vehicles and appropriate data handling are the essential features of a MaaS platform for the MSPs. For the travelers, although retrieval of personal information and enhancing services, such as suggestions on local events and concert tickets, are expected, they are considered less significant. Our study provides insights into the features of MaaS platforms through synthesized and prioritized features -including their relationships-which would be helpful both for research and practice.

A Basic Framework for Privacy Protection in Personalized Information Retrieval

Personalized information retrieval is an effective tool to solve the problem of information overload. Along with the rapid development of emerging network technologies such as cloud computing, however, network servers are becoming more and more untrusted, resulting in a serious threat to user privacy of personalized information retrieval. In this paper, we propose a basic framework for the comprehensive protection of all kinds of user privacy in personalized information retrieval. Its basic idea is to construct and submit a group of well-designed dummy requests together with each user request to the server, to mix up the user requests and then cover up the user privacy behind the requests. Also, the framework includes a privacy model and its implementation algorithm. Finally, theoretical analysis and experimental evaluation demonstrate that the framework can comprehensively improve the security of all kinds of user privacy, without compromising the availability of personalized information retrieval.

An empirical research and comparative analysis of clustering performance for processing categorical and numerical data extracts from social media

Social media has significantly influenced modern lifestyle and the way in which most of the industries operate their business. Social media data refers to the contents created by users during their social interactions in the form of text, sound, visuals, etc. It has now evolved as the major source of information for different industry verticals like retail, marketing, advertising, tourism, hospitality, education, etc. The huge volume of data resulted in the necessity for better and efficient procedures for personalized information retrieval. Traditional data mining and information retrieval techniques based on content-based and/or collaborative filtering proved computationally costly and less scalable against the volume it must deal with. Adoption of clustering techniques is a potential solution for this problem as it can minimize the amount of data required to be managed in industrial applications like recommender systems. This empirical research focuses on evaluating multiple clustering algorithms with the goal of finding an ideal solution for clustering numerical data extracted from social media sources. Three different publicly available datasets with varying number of attributes and records from tourism domain are used for the experiments conducted as part of this work

Personalized Information Retrieval from Friendship Strength of Social Media Comments

Social networks have become an important venue to express the feelings of their users on a large scale. People are intuitive to use social networks to express their feelings, discuss ideas, and invite folks to take suggestions. Every social media user has a circle of friends. The suggestions of these friends are considered important contributions. Users pay more attention to suggestions provided by their friends or close friends. However, as the content on the Internet increases day by day, user satisfaction decreases at the same rate due to unsatisfactory search results. In this regard, different recommender systems have been developed that recommend friends to add topics and many other things according to the seeker’s interests. The existing system provides a solution for personalized retrieval, but its accuracy is still a problem. In this work, we have proposed a personalized query recommendation system that utilizes Friendship Strength (FS) to recommend queries. For FS calculation, we have used the Facebook dataset comprising of more than 22k records taken from four different accounts. We have developed a ranking algorithm that provides ranking based on FS. Compared with existing systems, the proposed system can provide encouraging results. Key research groups and organizations can use this system for personalized information retrieval.

Proxy-Terms Based Query Obfuscation Technique for Private Web Search

Search engines store users’ queries in a query log for performing personalized information retrieval. However, query logs cause privacy concerns and reveal a lot of information about individuals if used against them. Private web search (PWS) provides a privacy-preserving information retrieval (IR) facility which allows users to retrieve information from an IR system without revealing true search queries. Current PWS techniques that are explored in the domain of web search are query obfuscation-based private web search (OB-PWS). These techniques achieve web privacy by injecting cover queries into the user profiles. However, existing OB-PWS techniques submit true queries along with cover queries and achieve query obfuscation in an isolated manner without considering the similarity between consecutive queries. In this article, we propose a proxy-terms based query obfuscation technique that allows users to retrieve information from an IR system through proxy queries without submitting true queries. IR system automatically generates cover queries and true queries from the proxy queries and cannot differentiate whether the user is trying to retrieve information for the cover queries or true query. We analyzed the effectiveness of the proposed technique on test queries, and develop a similarity metric for testing the accuracy of the proposed technique. Promising results of experiments confirm the effectiveness of the proposed technique.

Semantic Representation of a Geo-Social User Profile for a Personalised Information Retrieval

The user-centred information retrieval needs to introduce semantics into the user modelling for a meaningful representation of user interests. The semantic representation of the user interests helps to improve the identification of the user’s future cognitive needs. In this paper, we present a semantic-based approach for a personalised information retrieval. This approach is based on the design and the exploitation of a user profile to represent the user and his interests. In this user profile, we combine an ontological semantics issued from WordNet ontology, and a personal semantics issued from the different user interactions with the search system and with his social and situational contexts of his previous searches. The personal semantics considers the co-occurrence relations between relevant components of the user profile as semantic links. The user profile is used to improve two important phases of the information search process: (i) expansion of the initial user query and (ii) adaptation of the search results to the user interests.

The Protection of User Preference Privacy in Personalized Information Retrieval: Challenges and Overviews

Abstract This paper reviews a large number of research achievements relevant to user privacy protection in an untrusted network environment, and then analyzes and evaluates their application limitations in personalized information retrieval, to establish the conditional constraints that an effective approach for user preference privacy protection in personalized information retrieval should meet, thus providing a basic reference for the solution of this problem. First, based on the basic framework of a personalized information retrieval platform, we establish a complete set of constraints for user preference privacy protection in terms of security, usability, efficiency, and accuracy. Then, we comprehensively review the technical features for all kinds of popular methods for user privacy protection, and analyze their application limitations in personalized information retrieval, according to the constraints of preference privacy protection. The results show that personalized information retrieval has higher requirements for users’ privacy protection, i.e., it is required to comprehensively improve the security of users’ preference privacy on the untrusted server-side, under the precondition of not changing the platform, algorithm, efficiency, and accuracy of personalized information retrieval. However, all kinds of existing privacy methods still cannot meet the above requirements. This paper is an important study attempt to the problem of user preference privacy protection of personalized information retrieval, which can provide a basic reference and direction for the further study of the problem.

Privacy concerns in e‐commerce: A multilevel meta‐analysis

AbstractWhilst the rapid advancement of technology in the 21st century has facilitated the online collection, storage, retrieval, manipulation, and transmission of individuals' personal information, it has also led to a concomitant rise in privacy concerns amongst e‐commerce users. Although privacy concerns have received considerable attention in the e‐commerce literature, to date, empirical research has tended to report somewhat erratic and inconsistent findings in the context of consumer privacy. Accordingly, the relationships between the antecedents, privacy concerns, and the outcome variables in e‐commerce contexts remains unclear. To remedy such deficiencies in the literature, this study adopts the meta‐analytic approach to gather and make sense of the inconsistent and mixed empirical findings reported in the literature. The findings show that risk perceptions trigger privacy concerns while benefit perceptions, familiarity, reputation, privacy policy, and trust mitigate privacy concerns which in turn affect customer attitude and usage of e‐commerce platforms. To investigate the possible reasons for inconsistent findings, we performed a moderation analysis which suggests that methodological moderators, that is, type of article, research methods, and sample type; and contextual moderators, i.e. country where the study was carried out, and gender dominance in a sample set, can cause inconsistencies in the findings. Theoretically, this meta‐analysis contributes to the Antecedents; Privacy Concerns; Outcome variables (APCO) Model, and the literature on consumer privacy in the context of e‐commerce. Practically, the findings provide guidelines to e‐commerce businesses to effectively address customers' privacy concerns.

PAPIR: privacy-aware personalized information retrieval

The problem of information overload on the Internet increased the need for personalized information retrieval (PIR) systems capable of providing information that corresponds to the user interests. Although, for most people, the word personalization comes with trust issues and privacy concerns. Since giving the user a personalized browsing experience usually comes at the cost of his privacy. Thus, most people are afraid of using such applications. To address this issue, we propose a new model for privacy protection in PIR systems. Our model aims at achieving a trade-off between the personalization quality and the privacy risk, to keep the latter under control. We have studied the assets and drawbacks of the existing profile-based PIR structures, from a privacy protection perspective, along with the possible privacy threats in this field in a threat modeling approach. The model we propose is based on the vector space model and targets profile-based PIR systems. It uses query expansion and re-ranking algorithms on the client-side to ensure personalization quality. While privacy protection is ensured during the personalization process, by taking into consideration the user’s privacy requirements, and through encryption. We use the Advanced Encryption Standard (AES) algorithm to protect user data at-rest and a fully homomorphic encryption (FHE) scheme for data in-transit and in-use protection. To prove the feasibility and efficiency of our model, this paper includes a proof-of-concept implementation with proper experimental results.

Personalized Retrieval in the Medical Domain: A NoSQL Solution Based on Ontology Building

Managing medical information in a Big Data context is a challenging task since searching for relevant information in a large volume of data needs advanced treatments. Medical data is a special type of data because it comes from different sources and in different formats and encapsulates medical knowledge. Personalized retrieval is necessary when it comes to medical data management. In fact, the patient’s medical record needs to be taken into account in order to offer relevant documents since it contains his/her medical history. The proposed approach offers an ontology building process based on the patient’s medical record. The built ontology is then used for personalized information retrieval as well as user similarity computation. The approach is composed of three layers: (1) Data layer, (2) Treatment layer and (3) Semantic layer and offers three treatments: (1) Ontology building, (2) Query reformulation and (3) User similarity computation. An application supporting all three layers has been implemented and it allowed an experimental evaluation of the proposal. The results show an improvement in the relevancy of returned medical documents.

Out of sight and out of mind: Bookmarks are created but not used

Users commonly report problems in refinding important websites. To address this, some create Bookmarks (also called web Favorites) to improve the refinding of these websites. Previous research notes that these Bookmarks are rarely used. However, these prior studies did not systematically observe and quantify the frequency and success of Bookmark retrievals, or compare their success with other retrieval methods. The authors address those questions in this study using the elicited personal information retrieval technique, in which they asked their 50 participants to retrieve target URLs (uniform resource locators). Each participant received 21 targets, from which five were taken from the participants’ Bookmarks, and they were presented in random order to avoid raising suspicions. Although most of the participants created Bookmarks, they rarely used them for retrieval. Across all of the participants, only 41 (16%) of these 250 bookmarked retrieval targets were actually retrieved using the Bookmark facility. Of these 41 instances when Bookmarks were used, only 9 (4%) Bookmarks were retrieved using the Bookmarks menu hierarchy, while the remaining 32 were located in the browser’s upper bar, which was in full view of the participants. The results suggest that unless the Bookmarks were highly visible, the participants did not use them. Furthermore, for websites that the users had visited, bookmarked websites were not better retrieved than those that had not been bookmarked. The authors conclude by discussing possible explanations, as well as design and theory implications.

Collaboratively augmented UIP – Filtered RIP with relevancy mapping for personalization of web search

Personalized information retrieval has become more crucial and challenging in today’s time, as every user is demanding information according to their own perspectives. To achieve personalization, this paper proposes a methodology with a focus on the selection of an effective composition corresponding to various supporting modules of a personalization methodology. Collaborative tagging can be quite helpful in constructing User Interest Profile (UIP) and Resource Illustration Profile (RIP). The proposed methodology also focuses on UIP augmentation using multiple strategies; and a novel approach has also been designed to handle outlier tags which caused ambiguity in collective RIP. Even a good UIP and RIP alone cannot create an efficient personalization methodology; they also require a suitable mapping with user’s query requirement. Therefore, in the proposed methodology, the fuzzy satisfaction requirement-based novel mapping functions have been designed to measure query relevance score and user interest relevance score for a web resource. These scores have been further used to calculate the post-relevance score of a web resource after a suitable trade-off. Experiments using the del.icio.us dataset show that the proposed methodology has outperformed each and every baseline by a considerable margin.

SIB Literature Services: RESTful customizable search engines in biomedical literature, enriched with automatically mapped biomedical concepts.

Thanks to recent efforts by the text mining community, biocurators have now access to plenty of good tools and Web interfaces for identifying and visualizing biomedical entities in literature. Yet, many of these systems start with a PubMed query, which is limited by strong Boolean constraints. Some semantic search engines exploit entities for Information Retrieval, and/or deliver relevance-based ranked results. Yet, they are not designed for supporting a specific curation workflow, and allow very limited control on the search process. The Swiss Institute of Bioinformatics Literature Services (SIBiLS) provide personalized Information Retrieval in the biological literature. Indeed, SIBiLS allow fully customizable search in semantically enriched contents, based on keywords and/or mapped biomedical entities from a growing set of standardized and legacy vocabularies. The services have been used and favourably evaluated to assist the curation of genes and gene products, by delivering customized literature triage engines to different curation teams. SIBiLS (https://candy.hesge.ch/SIBiLS) are freely accessible via REST APIs and are ready to empower any curation workflow, built on modern technologies scalable with big data: MongoDB and Elasticsearch. They cover MEDLINE and PubMed Central Open Access enriched by nearly 2 billion of mapped biomedical entities, and are daily updated.

Probability model of sensitive similarity measures in information retrieval

In today’s Internet age, a lot of data is stored and used, which is very important. In people’s daily life, if these data are sorted, information retrieval technology will be used, and in information retrieval, some information retrieval inaccuracies often appear. Information retrieval model is an important framework and method for fast, complete, and accurate user information retrieval. With the rapid development of information technology, great changes have taken place in people’s production and life. Various information network technologies are widely used in people’s lives. The resulting flow of information shows explosive growth, information retrieval. User requirements are getting higher and higher. How to complete personalized information retrieval in a large amount of mixed information, so that retrieval technology can help us obtain effective retrieval results, has become a realistic problem worth exploring. In this article, the application of probability model based on sensitive similarity measure in information retrieval model is analyzed, and a similarity measure algorithm based on spectral clustering is proposed. By improving the similarity measure, the sensitivity problem of scale parameters is overcome and the retrieval precision is improved. In order to better reflect the superiority of the proposed algorithm, this article compares with ng-jordan-weiss (NJW) and deep sparse subspace clustering (DSSC) algorithms. The experimental results show that the proposed algorithm is superior to NJW and DSSC algorithms for different data sets in different evaluation indicators (Rand and F-measure).