Journal of Information & Knowledge Management

Domain Knowledge Graphs are becoming increasingly significant in a multitude of sectors, as they collate distinctive data from a plethora of disciplines. However, in some domains, generic knowledge representation is limited, resulting in problems like duplication and knowledge loss. To solve the above issues, the study first builds a Knowledge Representation Learning model for training and then constructs a Domain Knowledge Graph inference algorithm based on Knowledge Representation Learning for knowledge inference and quality evaluation. The results indicated that the effectiveness of the knowledge-inference-raised method was the best in datasets of different sizes, with the average accuracy, Hits@1, Hits@3 and Hits@10 of 0.906, 0.914, 0.942 and 0.948, respectively. In the results of the link prediction task, the study method had a poor performance for only one relation, with an average accuracy rate of 0.675. In the application results of knowledge graph quality assessment, the interpretability and data fusion ability of both Domain Knowledge Graphs were strong, with the accuracy indexes exceeding 0.95, and the various indexes of simplicity and completeness exceeding 0.88. The above results indicate that our research method is effective in filling in the missing knowledge through knowledge reasoning and ensures data reliability.

The purpose of this study was to provide a holistic approach to establishing dyadic relationships between information sharing and operational performance as well as between operational and organisational performances into an extended relationship between ultimate suppliers and ultimate customers from the suppliers’ point of view. We introduced mediating impacts of information usage between information sharing and operational performance and perceived logistics performance between operational and organisational performances. The empirical data obtained from 209 retailers via a survey reveals that the integrated relationships are valid and significant. The mediating impacts of both information usage and perceived performance of logistics play a significant role in performance relationships and are paramount to the success of organisations.

Legal outcome prediction or legal judgement prediction is the process of making predictions of the possible outcome of a legal proceeding based on its case file. This work focuses on the importance of allowing less privileged or marginalised groups to have better access to law services. This work discusses the implementation of 13 different deep learning models and the results obtained for each of them. The range of models employed encompasses a spectrum from basic perceptron models to more complex convolutional network models and encoder-based models. The models like multi-layer perceptron (MLP), convolutional neural network (CNN) model, BERT, LegalBERT, DistilBERT, RoBERTa and the different combinations of these encoder-based models are implemented. The input to these models is the case files and their corresponding judgement status. Four different classes of petitions and appeals, submitted to the court, were considered to train all the models. The efficiency of these models was tested in terms of various performance metrics like training accuracy, testing accuracy, training loss, validation loss and $F$1 score.

Recently, personal health records (PHR) have been found to come into view as a patient-centric approach for exchanging health data. With the use of a PHR service, a physician may create, access and maintain all of their personal health information online in one convenient location. Additionally, it has greatly improved the efficient exchange, retrieval and storage of patient data. Each patient is also promised total ownership over their medical records and the option to share their data with a wider range of clients together with friends, family members or healthcare providers. Owing to the higher price of maintaining and building specific data centres, numerous PHR services were contracted out to third-party service providers. Therefore, this work is concerned with carrying out the secured sharing of PHR by inducing the signcryption model with an optimisation concept. In the signcryption model, the digital signature encryption is done using an improved AES approach. Further, the key generation is carried out using a self-customised shark smell algorithm (SC-SSA) algorithm. Finally, the evaluation is done to verify the improvement of the system that was used.

Most deep learning models face difficulties in analysing image information due to the concept of information bottlenecks and their corresponding methodologies. But, the information bottleneck is used for discarding redundant data and trying to maximise in favour of data directly relevant to the task-oriented information. However, managing information bottlenecks is challenging in the learning model process. Although convolutional neural networks are designed for small-scale processing, their inductive bias makes it difficult to learn contextual features. Thus, we have considered the theoretical learning model to justify the advantages of information bottleneck in deep learning model. We tried to use a fundamental information bottleneck in the vision transformer model. The channel density module cleans up task-related data, while the collected image representations are encouraged to be diverse through local connections in cumulative local transformer blocks. We considered the encoder and decoder methods that analyse the information bottleneck techniques in the deep learning model. This paper presents a rigorous learning theory that mathematically links information bottlenecks for generalisation errors, demonstrating the usefulness of information bottlenecks in deep learning. Our approach suggests that limiting information bottlenecks is crucial for managing errors in deep learning techniques. We conducted experiments across various mathematical models and learning environments to test the validity of our new mathematical insights. In many cases, generalisation errors correspond to unwanted information at hidden levels. We have considered boundary approaches using various scaling parameters and dimensions for the degree of information bottleneck. As per the estimation loss and error by different correlation approaches using generalisation gap methods, we found Spearman correlation having loss (0.86) and error (0.758), whereas Pearson correlation having loss (0.85) and error (0.76), respectively. We also considered outputs for model compression metrics and analysed them through comparative performance.

The purpose of this study is to address the issue of low efficiency and time-consuming monitoring of student classroom behaviour in teaching classroom scenarios, as well as how to apply deep learning to student classroom behaviour detection. Classroom teaching is the main way for students to acquire and master knowledge, and teachers need to pay attention to students’ behaviour in the classroom and make timely corrections. In this study, the teaching classroom of undergraduate students was used as an example for application and analysis. Based on the You Only Look Once v5s network, a student behaviour detection algorithm was designed by combining the ghosting module and optimised feature fusion structure, and a student behaviour association technology suitable for continuous teaching classrooms was designed. Through the visualisation results, it can be seen that the research method can accurately detect the behaviour categories of students, and the effect is better in target dense scenes. Through quantitative analysis using correlation technology, it can be concluded that there are very few students with the lowest level of seriousness, while the majority of students have a level of seriousness above the qualified level. The overall situation of the teaching classroom is good. These student behaviour analyses can energy the level of seriousness of students, form effective and objective classroom feedback, provide teachers with objective and accurate reference value, and promote the construction of smart classrooms.

Background: With the rapid development of information technology, natural language processing has become an important branch, which has received widespread attention and application. The rise of this technology has brought revolutionary changes to the field of education, especially in achieving the Sustainable Development Goal (SDG) of quality education. In the field of online English education, Multiple Choice Questions (MCQs) are a common assessment method, and their accuracy directly affects students’ learning outcomes and feedback quality. Purpose: In order to improve the accuracy and effectiveness of MCQs, an innovative search method based on fuzzy tree matching is proposed. Methodology: This method effectively captures the core features of grammar problems and calculates the similarity between problems by introducing the Parse-key tree structure. The Parse-key tree consists of subtrees and positional information. The subtree structure reduces the impact of noise on the target grammar structure information and inserts spatial position information to determine grammar knowledge points. Findings: The experimental results showed that in incomplete queries, the Mean Reciprocal Rank (MRR) value of the grammar MCQs problem was improved by 7.9% compared with the existing methods. When the recall rate was 0.1, the precision of our method remained above 0.4, while that of other methods was below 0.4. In complete queries, the MRR value of the proposed method was 29.6% higher than that of the traditional part of speech-sorting algorithms, and the precision usually remained between 0.2 and 1. However, when other methods reached 0.2, the precision of our method dropped below 0.2, demonstrating significant performance advantages. Based on the above results, the research method not only improves the efficiency and quality of knowledge acquisition, but also optimises the organisational structure of knowledge, enhances the sharing ability of knowledge, promotes innovation and application of knowledge and provides new ideas and technical support for the development of online education. Contribution: This research proposes an innovative search method based on fuzzy tree matching, which can effectively improve the accuracy and effectiveness of MCQs assessment in online English education.

In the era of digital transformation, the popularity of artificial intelligence technology has driven the development of the field of Human Resource Management (HRM) in enterprises, and data-driven HRM has become the key to the development and success of enterprises. Due to the late start, data-driven HRM is still in the development stage in China. To address the problem that traditional HRM lacks accuracy and real-time, this study proposes a Growth Potential Prediction (GPP) model based on the combination of all class features and a Random Forest (RF) classifier. The results of the study showed that the prediction accuracy of interpersonal background, work ability, educational status and organisational environment characteristics under the screening index based on $F$1 value were 85.01%, 66.34%, 66.41%, 44.19% and 64.27%, respectively. The accuracy of ROS under the RF classifier and the macro $F$1 value were 87.31% and 87.24%. After adding the network features, the accuracy and macro $F$1 values of the feature ensemble could reach 90%, and the accuracy and macro $F$1 values of all class feature combinations were 89.01% and 89.10%, respectively. In summary, the GPP model based on the combination of all class features and the RF classifier proposed in the study can accurately predict the growth potential of employees and provide some guarantee for mining the growth pattern of employees.

With the continuous reform of education informatization, modern information technology gradually became a key technology in education. An intelligent question-answering system was constructed in this research for knowledge management services. It included an entity recognition model based on bidirectional long short-term memory attention conditional random field and a question classification model based on a bidirectional encoder using Transformers text convolutional neural network. The entity recognition model added a reverse long short-term memory propagation structure, while introducing an attention mechanism and conditional random field model. The question classification model utilized a preprocessed structure to output a feature matrix with richer semantic information, while utilizing a bidirectional encoder to extract local features of short text sentences. These experiments confirmed that the constructed model had a high precision of 0.945, which was much higher than other models. The recall of this proposed model was 0.077 higher than the control models. For precision, the proposed model had good recognition performance for methods, files and modules, with values of 0.9040, 0.9050 and 0.9160, respectively. For the recall, the question classification model had the best recognition effect on files, at 0.9060. For the $F$1-score, the proposed model had the best recognition performance for modules, at 0.9100. The named entity recognition model constructed had high accuracy and $F$1-score, reaching 0.931 and 0.923, indicating that it had the best classification performance in the dataset. Through intelligent question-answering systems, automatic answering of questions can be achieved, effectively improving knowledge services. This research has integrated knowledge and technology from multiple disciplines such as computer science, artificial intelligence, and natural language processing to promote the cross-integration of these disciplines in knowledge management. This interdisciplinary method not only broadens the research perspective of knowledge management but also provides new ideas and methods for the development of related fields.

Teaching in the modern academic environment is more complex, which makes it necessary to examine the challenges Kerala’s self-finance institutions arts and science institutes’ staff members are experiencing. This study investigates workplace stress among academic staff in self-financing arts and science institutes in Kerala. It analyses stress management techniques, including wellbeing initiatives and emotional intelligence training, and identifies the main stressors based on knowledge management (KM). The goal of the study is to improve faculty members’ performance and wellbeing in these institutions by providing information regarding relevant stress-reduction techniques that are specific to KM. The factors causing workplace stress have been measured as (a) demographic factors, (b) institutional factors and (c) psychological factors. Also, the stress management activities for the faculty members were considered a wellbeing programme, emotional intelligence, training and development. Initially, data collection was performed and 156 faculty members worked in arts and science self-financing colleges of the Kerala district. For a survey, in accordance with the proposed hypothesis, a systematic questionnaire has been created and sent to the faculty members who operate in Kerala. Moreover, statistical tests were performed on the gathered data using structural equation modelling (SEM) techniques. This investigation has revealed the characteristics that affect the faculty and staff working for an organisation in terms of workplace stress and excellent fit. The findings reveal that demographic demographics, institutional issues and stress management strategies influence faculty members in Kerala’s arts and self-finance colleges.

The core of artificial intelligence (AI) research is intelligent information processing, and fuzzy reasoning and fuzzy neural networks are significant research areas in models of intelligent information processing. The study examines the robustness of fuzzy reasoning by using the property characteristics of triangular modal operators and implication operators, which are used to analyse the Lipschitz property of fuzzy operators and its corresponding property characteristics of triangular modal operators and implication operators. Inference operators are used to investigate the robustness and error-control capabilities of fuzzy operators for fuzzy reasoning. The results showed that the robustness of quasi-copula fuzzy reasoning with fusion rules is better, and the corresponding maximum output perturbation is lowest at 0.36 when the fuzzy operator is T_M&R_MTc and the corresponding maximum output error is lowest at 0.41 when the fuzzy operators are T_M&R_MKl and T_M&R_Ma. This shows the optimal performance of the fuzzy associative memory model with fuzzy rule for quasi-copula fuzzy operators and can improve the theoretical technology. The best-performance fuzzy copula operator of the fuzzy associative memory model can enhance the theoretical foundation and technical support for the processing of intelligent information.

This research aims to evaluate the extent to which a specific set of factors affect patients’ intention to adopt and use smart health applications. These factors, derived from the unified theory of acceptance and use of technology (UTAUT), include technology anxiety (TA), perceived reliability (PR), inertia (Int) and privacy concern (PC). Data were collected from 259 participants via a cross-sectional survey. The researchers employed the partial least squares (PLS) method for data analysis using SmartPLS, a statistical tool based on structural equation modelling (SEM). The findings indicate that performance expectancy, facilitating conditions, Int and PR significantly influence patients’ intentions to adopt smart health services. However, no significant relationships were found between effort expectancy, social influence, PC, TA and the intention to adopt smart health services. This study contributes to the literature by integrating additional factors (TA, PR, Int and PC) into the UTAUT framework. Limitations of the study include the cross-sectional design, which may not capture changes over time, and the reliance on self-reported data, which may be subject to biases. Future research should consider longitudinal studies to examine the evolving nature of technology adoption and explore additional contextual factors that might influence patient intentions. This research provides practical insights for healthcare practitioners aiming to enhance the adoption of smart health applications.

With the change in the communication pattern, end-users are engaging in creating content and referring to the content created by other users while making purchase decisions. This research aims at modelling factors affecting consumers’ usage intention (UI) towards user-generated content (UGC) using Need for Cognition (NfC) as a moderator of the proposed relationships. The factors affecting consumers’ UI involve perceived usefulness (PU), source credibility (SC), information quality (IQ) and NfC. Further, a novel attempt has been made by using the neural network approach to assess the predictive accuracy of the model. A structured questionnaire was used to collect data from 298 consumers through a survey. Data were analysed using two-stage structural equation modelling (SEM) and artificial neural network (ANN). All the independent variables viz., PU, SC, IQ and NfC significantly affect attitude towards UGC, which in turn affects UI. Results of multi-group analysis and a series of chi-square difference tests reveal that a NfC significantly moderates the relationship between PU and attitude, as well as that between SC and attitude. The root mean square error values from the neural network analysis suggest that the models show good predictive accuracy. This study provides a novel assessment of the usage of a hybrid SEM-ANN approach for understanding of UGC by incorporating NfC as a moderator in shaping consumers’ attitudes and intentions to use UGC.

To meet the needs of our knowledge-based society, students must have problem-solving, critical thinking, teamwork, learning-to-learn, and communication skills. Transferring knowledge is important to using these skills in non-school environments, communicating effectively, and applying content in new situations and future work. Flipped Learning has the potential to promote Deep Learning (DL). This paper describes a case study involving Flipped learning, to engage students in DL in an Economic intelligence course. We use the lens of Cultural-Historical Activity Theory (CHAT) to understand how the relationship between the human mind (what people think and feel) and activity (what people do) operates in mixing Flipped classrooms with mind-mapping techniques to promote DL. The evaluation of the results shows that students have a deeper understanding of the subject, but it also confirms a strong need for time and generalisation. The new pedagogy has to be introduced earlier in the school and integrated within the curricula to reinforce the students’ capacity for transfer and application. Isolated experimentation by teachers of Flipped Classrooms and Mind Mapping is insufficient to transfer knowledge in problem-solving situations. This limitation could negatively impact future entrepreneurs’ cognitive skills and task identity. Entrepreneurs as the main drivers of economic development need competencies in information and knowledge searching to reduce environmental uncertainty. They also have to make sense of early warning information to convert it into knowledge supporting the decision-making process.

Our research is centred on drug knowledge discovery, involving the integration of drug knowledge graphs with machine learning techniques. To address the challenges of burdensome workload and inadequate classification accuracy associated with constructing individual models for each Adverse Drug Reaction (ADR), we developed and refined a Convolutional Neural Network (CNN) based on knowledge graph embedding (KGE) and deep learning methodologies. The outcomes of our study demonstrate that our proposed predictive model achieves remarkable precision and reliability, enabling comprehensive exploration of drug-reaction relationships. Our contributions introduce innovative models and techniques that drive the advancement of intelligent healthcare and demonstrate the scalability of data science applications in the medical domain.

Smart farming is an enhanced option for increasing food production, resource management and labour. Existing prediction methods need trained experts to analyse the data, which is a time- consuming process, and thus, there is a need for smart farming with the Internet of Things (IoT). Hence, a well-developed IoT-assisted smart agriculture model is proposed for managing agricultural needs to improve the economic value of farmers. The proposed framework constitutes four major aspects like Crop prediction, Crop yield prediction, Plant disease prediction and Smart irrigation. Initially, the crop images are taken as the input and fed to the Multi-scale Adaptive and Attention-based Convolution Neural Network with Atrous Spatial Pyramid Pooling (MACNN-ASPP), where the dissimilar crops are classified and obtained. In the second aspect, the crop image as well as crop-related data, are fetched from the data sources and given as input. Further, the deep features of the image are extracted that are added with soil and environment condition data. Then this feature is subjected to the Multi-scale Adaptive and Attention-based One-Dimensional Convolution Neural Network with Atrous Spatial Pyramid Pooling (MA1DCNN-ASPP) for crop yield prediction. While in the third aspect, the leaf images are assembled from the data file and fed as input to the MACNN-ASPP for detecting the variety of diseases affecting the plants. In the final aspect, smart irrigation is done by collecting field images with related data. Further, the deep features retrieved from the field images are applied to MA1DCNN-ASPP, where the different conditions of the field will be attained. In order to develop the model adaptively, the hyper-parameters in the network are optimised using the Improved Reptile Search Algorithm (IRSA). Finally, the investigation is done over the proposed methodology using multiple evaluation metrics. In contrast with other approaches, the proposed smart agriculture outperforms the performance of managing crops without any hazardous effects. Accuracy validation executed in the implemented IRSA-MACNN-ASSP-based crop yield prediction model accomplished better efficiency as 6.89%, 4.45%, 2.19% and 1.08% better than the classical techniques like CNN, LSTM, 1DCNN and MA1DCNN-ASPP.

In today’s data-driven world, the volume of information produced daily is staggering. Without a robust data engineering strategy, companies face the risk of prolonged delays, decreased productivity, dissatisfied customers, and strained business relationships. Effective data management and data modelling are critical for transforming this vast amount of information into valuable insights that drive business growth and provide a competitive edge. By gathering and analysing data through these methods, businesses can make informed decisions that significantly impact their growth and success. Data modelling and data management are both critical components of working with data, but they focus on different aspects of handling and utilising data within an organisation. Understanding the distinction between these two areas is crucial for effectively managing data within an organisation and ensuring that data systems are well-designed and properly maintained. By delving into the specifics of data modelling and data management, this paper aims to provide a comprehensive understanding of how these practices can be leveraged to enhance organisational efficiency, productivity, and decision-making. We provide a comprehensive and insightful exploration of data modelling and data management while highlighting their critical roles in modern business environments.