Journal of Information & Knowledge Management

Diabetic retinopathy is a leading cause of blindness among diabetic patients, and early detection is crucial. This research proposes DiabNet, a novel convolutional neural network (CNN) architecture designed to enhance the accuracy, efficiency, and robustness of diabetic retinopathy detection from retinal images. DiabNet incorporates unique features like skip connections, attention mechanisms, and batch normalisation to improve feature extraction. The paper details DiabNet’s architecture, feature extraction, and training process. Evaluation on a standard dataset shows that DiabNet surpasses existing methods in accuracy, efficiency, and robustness. The research also explores the interpretability of DiabNet and suggests future research directions. The potential impact of DiabNet includes improved early detection and management of diabetic retinopathy. In addition, DiabNet’s deployment as a mobile app enables convenient and accessible diabetic retinopathy screening. Finally, it is noted that DiabNet, as a mobile app, has the potential to significantly impact the field of diabetic retinopathy detection, leading to improved early detection of diabetic retinopathy. The experimental validation proves that the proposed DiabNet architecture is feasible for real-time deployment yielding an accuracy of 98.72%.

Cloud is a model of computing providing on-demand availability of computer system resources, mainly data storage and computing power, without the user’s active management. This provides sharing and reduces user computing and storage costs. With the development of cloud scale and intensification, cloud security is becoming a vital issue in the cloud computing field. Here, a newly optimised hybridisation algorithm, namely, Shuffled Shepherd Fire Hawk Optimization (SSFHO) algorithm is proposed for key generation to initiate privacy protection in the cloud. The different entities involved in this approach are Data Owner (DO), Data User (DU), blockchain, and Cloud Service Provider (CSP). Here, the access control process includes seven categories, such as initialisation, registration, generation of key, data access control, authorisation, authorisation revocation and data protection. Also, registration is done in three phases, like registration of cloud, registration of user and resource publishing. Moreover, SSFHO is the integration among Shuffled Shepherd Optimization Algorithm (SSOA) and Fire Hawk Optimization (FHO) algorithm. The performance of the model is done with various performance metrics like, authorisation time, privacy rate, and memory. These revealed high values of privacy rate of 0.93, low values of authorisation time and memory of 0.559 and 0.843 GHz.

This research examines the relationship between sustainable m-Gov services and citizen satisfaction with m-Gov services. A multidimensional conceptualisation of sustainable m-Gov services is defined to examine citizen satisfaction. The research model was tested using empirical data collected from 687 m-Gov service users through PLS-SEM. The results showed that service availability, contact, responsiveness, efficiency, and privacy significantly influenced m-Gov service quality. Mobile self-efficacy, perceived trust, and perceived functional benefit are critical for m-Gov adoption. However, perceived compatibility and perceived ability-to-use did not explain the m-Gov adoption. The findings of m-Gov service quality and m-Gov adoption interactions supported their role in predicting sustainable m-Gov services, thereby increasing citizen satisfaction. The outcome of this study is vital for government strategies, public administration, policymakers, and government service delivery literature and provides citizen-centric m-Gov services. Thus, the government and citizens adopting m-Gov services can benefit from the tested model towards increasing the sustainable offering of m-Gov services.

To build a comprehensive learning resource system covering learning process, learning users, learning resources and other aspects, an intelligent learning system for innovation and entrepreneurship in universities was constructed and designed combined with knowledge visualisation theory. The system includes student information record module, system recommendation calculation module, learning material module, and knowledge visualisation module. The research focuses on the system recommendation calculation module, including the User-based Collaborative Filtering Algorithm (UCF) and the Joint Recommendation (JR) algorithm of the application content recommendation algorithm. The best neighbour number of JR is 20, and the best transmission path length is 2. The recommendation effect is optimal under this setting. The convergence iterations of the training and test set of the joint recommendation algorithm are 80 times and 100 times, respectively. For the training set and test set, the accuracy, precision, sensitivity, recall, running time, error, and other performance evaluation indicators of the joint recommendation algorithm are better than the corresponding values of other recommendation algorithms. The system has a good design effect and experience effect in the application of the online learning platform.

In order to strengthen the evaluation of the quality of sustainable entrepreneurship education in universities, a study was conducted to calculate the correlation degree of the quality of sustainable entrepreneurship education in universities based on the grey correlation algorithm. The study also constructs an evaluation index system and evaluation model from three aspects: the foundation of the entrepreneurship education environment, the allocation of entrepreneurship education resources and the performance of starting an undertaking education result. The result that the evaluation model is more similar to the actual situation indicates that the evaluation indexes have a certain degree of scientificity. The evaluation method of starting an undertaking sustainable education in C&U based on the grey correlation degree algorithm has higher accuracy, recall, and F1 value compared with the traditional way. Therefore, the method has a better evaluation effect for the quality evaluation of entrepreneurship sustainable education in C&U.

Recent advancements in Natural Language Processing (NLP) have made sentiment analysis an essential component of a variety of NLP jobs, including recommendation systems, question answering, and business intelligence products. While sentiment analysis research has, to put it mildly, been widely pursued in English, Telugu has barely ever attempted the task. The majority of research works concentrate on analysing the sentiments of Tweets, news, or reviews containing Hindi and English words. There is a growing interest among academics in studying how people express their thoughts and views in Indian languages like Bengali, Telugu, Malayalam, Tamil and so on. Due to a paucity of labelled datasets, microscopic investigation on Indian languages has been published to our knowledge. This work suggested a sentence-level sentiment analysis on multi-domain datasets that has been collected in Telugu. Deep learning models have been used in this work because it demonstrates the significant expertise in sentiment analysis and is widely regarded as the cutting-edge model in Telugu Sentiment Analysis. Our proposed work investigates a productive Bidirectional Long Short-Term Memory (BiLSTM) Network and Bidirectional GRU Network (BiGRU) for improving Telugu Sentiment Analysis by encapsulating contextual information from Telugu feature sequences using Forward-Backward encapsulation. Further, the model has been deployed by merging the domains so as to predict the accuracy and other performance metrics. The experimental test findings show that the deep learning models outperform when compared with the baseline traditional ML methods in four benchmark sentiment analysis datasets. There is evidence that the proposed sentiment analysis method has improved precision, recall, F1-score and accuracy in certain cases. The proposed model has achieved the F1-score of 86% for song datasets when compared with the other existing models.

By the end of 2021, the worldwide cryptocurrency market valuation had hit a whole high of $3 trillion. Blockchain technology underpins cryptocurrencies such as Bitcoin and Ethereum. The adoption of blockchain, as well as the technology and products it enables, will continue to have a significant influence on company operations. However, blockchain technology is much more than a secure cryptocurrency transfer method. It may be utilised in areas other than finance, including as healthcare, insurance, voting, welfare benefits, gaming and artist royalties. The global economy is prepared for the blockchain revolution, with the technology already having an influence on business and society on many levels. If the term “revolution” seems extreme, consider that eight of the world’s ten largest corporations are developing a variety of blockchain-based solutions. Any enterprise or organisation that is engaged in the recording and oversight of any type of transaction stands to profit from shifting its operations to a blockchain-based platform. This paper discusses about the various roles blockchain has taken up over the years in several different industries along with the future opportunities and scope of expansion to numerous other professional sectors.

As a result of late detection and resource limitations during any software evaluation, there have been several software-related breakdowns or malfunctions. Many people have begun focussing on the test cases or alternatively the priority of validation suites after identifying the difficulties in the regression testing process of any product. The test case prioritisation technique is presented as a solution to this problem. It increases the fault detection rate. Earlier research studies have been implemented many techniques, but the rate of fault detection is not up to the mark. To overcome this drawback, we proposed HOTCP (Hybrid Optimal Test Case Prioritisation with Multi-Objective Constraints) model, which includes two steps: first is test case generation and the next one is test case prioritisation. The test case is generated from the released software. Consequently, test case prioritisation will be done by the optimisation strategy, in which the multi-objective function will be defined based on the constraints like statement coverage, branch coverage, contribution index and fault exposing potential. For this optimisation process, a new algorithm is proposed termed as CCCOA (Customised Coot and Chimp Optimisation Algorithm). The COOT optimisation and Chimp optimisation algorithms are combined in this algorithm. The system produces prioritised test cases, and the performance of the proposed method is validated in comparison to the traditional methods in terms of several metrics.

With the help of advancements in connected technologies, social media and networking have made a wide open platform to share information via audio, video, text, etc. Due to the invention of smartphones, video contents are being manipulated day-by-day. Videos contain sensitive or personal information which are forged for one’s own self pleasures or threatening for money. Video falsification identification plays a most prominent role in case of digital forensics. This paper aims to provide a comprehensive survey on various problems in video falsification, deep learning models utilised for detecting the forgery. This survey provides a deep understanding of various algorithms implemented by various authors and their advantages, limitations thereby providing an insight for future researchers.

Text continues to remain a relevant form of representation for information. Text documents are created either in digital native platforms or through the conversion of other media files such as images and speech. While the digital native text is invariably obtained through physical or virtual keyboards, technologies such as OCR and speech recognition are utilised to transform the images and speech signals into text content. All these variety of mechanisms of text generation also introduce errors into the captured text. This project aims at analysing different kinds of errors that occur in text documents. The work employs two of the advanced deep neural network-based language models, namely, BART and MarianMT, to rectify the anomalies present in the text. Transfer learning of these models with available dataset is performed to finetune their capacity for error correction. A comparative study is conducted to investigate the effectiveness of these models in handling each of the defined error categories. It is observed that while both models can bring down the erroneous sentences by 20+%, BART can handle spelling errors far better (24.6%) than grammatical errors (8.8%).

Population has increased drastically over the years and new diseases compete with the population. Immunisation is a preventive measure, which makes the person resistant or immune to the disease. Vaccination stimulates our own immune system against infection or diseases. Vaccines are available for more than twenty life-threatening diseases and it saves millions of lives throughout the world. In the 70th World Assembly conducted in the year 2017, around 194 countries participated and took the oath to strengthen vaccination thereby achieving goals of Global Vaccine Action Plan (GVAP). In spite of remarkable immunisation progress, approximately 20 million infants are not exposed to vaccination every year. The immunisation progress has stalled or even reversed in some countries, and there is a real risk that complacency will undermine past achievements. This paper considered the database of vaccine consumption rate from many countries issued by WHO to analyse the reason for poor access to vaccine with respect to the morality and poverty levels. For this analysis, the relation between vaccine consumption by children below five years, the children’s death rate record issued by the United Nations Children’s Education Fund (UNICEF), poverty index issued by the United Nations are considered. Multivariate linear regression algorithm is used to identify the correlation between datasets. The result shows that an increase in vaccination coverage results in reduction on mortality rate in most of the countries. A correlation coefficient of 0.7 was found between IMR and the vaccine dosages. Sub-Sahara countries’ poverty index has direct impact on the declined view of vaccination coverage.

Collaborative filtering-based recommendation systems have become significant in various domains due to their ability to provide personalised recommendations. In e-commerce, these systems analyse the browsing history and purchase patterns of users to recommend items. In the entertainment industry, collaborative filtering helps platforms like Netflix and Spotify recommend movies, shows and songs based on users’ past preferences and ratings. This technology also finds significance in online education, where it assists in suggesting relevant courses and learning materials based on a user’s interests and previous learning behaviour. Even though much research has been done in this domain, the problems of sparsity and scalability in collaborative filtering still exist. Data sparsity refers to too few preferences of users on items, and hence it would be difficult to understand users’ preferences. Recommendation systems must keep users engaged with fast responses, and hence there is a challenge in handling large data as these days it is growing quickly. Sparsity affects the recommendation accuracy, while scalability influences the complexity of processing the recommendations. The motivation behind the paper is to design an efficient algorithm to address the sparsity and scalability problems, which in turn provide a better user experience and increased user satisfaction. This paper proposes two separate, novel approaches that deal with both problems. In the first approach, an improved autoencoder is used to address sparsity, and later, its outcome is processed in a parallel and distributed manner using a MapReduce-based $k$-means clustering algorithm with the Elbow method. Since the $k$-means clustering technique uses a predetermined number of clusters, it may not improve accuracy. So, the elbow method identifies the optimal number of clusters for the $k$-means algorithm. In the second approach, a MapReduce-based Gaussian Mixture Model (GMM) with Expectation-Maximization (EM) is proposed to handle large volumes of sparse data. Both the proposed algorithms are implemented using MovieLens 20M and Netflix movie recommendation datasets to generate movie recommendations and are compared with the other state-of-the-art approaches. For comparison, metrics like RMSE, MAE, precision, recall, and F-score are used. The outcomes demonstrate that the second proposed strategy outperformed state-of-the-art approaches.

In WSN, extending the network life is still a major problem that needs to be tackled. Cross-layer protocols are used to get around these problems. In this paper, a new cross-layer design routing model is presented using a clustering-based technique. The proposed model is proceeding with the optimal cluster-based routing model via a new algorithm. Initially during the network generation, the node’s energy is predicted by the deep learning model termed as DNN model on the basis of the distance between node and sink as the input. Subsequently, during the clustering process, the cluster head is optimally selected via a new optimisation algorithm named Self-Improved Shuffle Shepherd Optimisation (SISSO) Algorithm. The cluster head selection is done by considering the constraints including Link quality, Distance, Overhead, Energy and Delay. Finally, a Modified Kernel Least Mean Square (MKLMS)-based data aggregation process is to eliminate the redundant data transmission. The performance of the SISSO method is proven superior over other conventional approaches with regard to the alive node and network lifetime. In the alive node analysis of supernode, the proposed SISSO model achieves the maximal number of alive supernodes at 2,000 rounds (i.e. 0.67) than other conventional methods.

This methodology combines geospatial clustering and Natural Language Processing (NLP) to create a framework for discovering unexplored geotags in social media. The framework contains the collection of data from social media platforms, the preprocessing of data with Pandas, Natural Language Toolkit (NLTK) and SpaCy libraries for the NLP analysis as well as for sentiment analysis and named entity recognition, followed by spatial clustering with Density-based Space Clustering of Noise Applications (DBSCAN), K-Means and HDBSCAN algorithms, then visualising with Matplotlib and Folium libraries. The data analysis and statistics were done using Pandas and NumPy libraries, and exploration through the selection and collection of more data based on the previous step. In addition, a prediction model has been developed to predict a location cluster using its name by comparing it to the preprocessed comma-separated values data file. Currently, there are certain locations like small-scale hospitals or unknown tourist places which are not currently tagged on available maps applications. This framework can be useful for researchers and policy makers to identify those locations and gain insights from social media data and find its potential for decision-making in various fields.

Conversational flow refers to the progression of a conversation, encompassing the arrangement of topics discussed and how responses are delivered. A smooth flow involves participants taking turns to speak and respond naturally and intuitively. Conversely, a more disjointed flow may entail prolonged pauses or difficulties establishing common ground. Numerous factors influence conversation flow, including the personalities of those involved, their familiarity with each other, and the contextual setting. A conversational graph pattern outlines how a conversation typically unfolds or the underlying structure it adheres to. It involves combining different sentence types, the sequential order of topics discussed, and the roles played by different individuals. Predicting subsequent sentences relies on predefined patterns, the context derived from prior conversation flow in the data, and the trained system. The accuracy of sentence predictions varies based on the probability of identifying sentences that fit the subsequent pattern. We employ the Graph Recurrent Neural Network with Attention (GRNNA) model to generate conversational graphs and perform next-sentence prediction. This model constructs a conversational graph using an adjacency matrix, node features (sentences), and edge features (semantic similarity between the sentences). The proposed approach leverages attention mechanisms, recurrent updates, and information aggregation from neighbouring nodes to predict the next node (sentence). The model achieves enhanced predictive capabilities by updating node representations through multiple iterations of message passing and recurrent updates. Experimental results using the conversation dataset demonstrate that the GRNNA model surpasses the Graph Neural Network (GNN) model in next-sentence prediction, achieving an impressive accuracy of 98.89%.

Cybersecurity is becoming increasingly important with the rise in Internet usage. The two most frequent cyberattacks that can seriously harm a website or a server and render them inaccessible to other customers are denial of service (DoS) and distributed denial of service (DDoS) attacks. These attacks are so common and take many different forms, it is difficult to identify and respond to them with previous methods. Furthermore, computational complexity, inconsistency, and irrelevant data are problems for traditional intrusion detection methods. As a result, a powerful deep learning-based technique is applied in this study for the identification and categorisation of DoS and DDoS attacks. Refined Attention Pyramid Network (RAPNet)-based feature extraction is used in this proposed framework to extract features from the input data. Then, Binary Pigeon Optimisation Algorithm (BPOA) is used to determine the best features. After choosing optimal characteristics, Densenet201-based deep learning is deployed to categorise the assaults in Bot-IoT, CICIDS2017, and CICIDS2019 datasets. Furthermore, the Conditional Generative Adversarial Network (CGAN) is used to provide extra data samples for minority classes to address the issue of imbalanced data. The findings show that the proposed model can precisely identify and categorise DoS and DDoS assaults in comparison to the existing intrusion detection approaches with 99.43%, 99.26%, and 99.38% accuracy for CICIDS2019, CICIDS2017, and BoT-IoT correspondingly.

Software bug prediction is mainly used for testing and code inspection. So, software bug prediction is carried out by network measures over the decades. But, the classical fault prediction method failed to obtain the semantic difference among various programs. Thus, it degrades the working of the prediction model, which is designed using these aspects. It is necessary to obtain the semantic difference for designing the prediction model accurately and effectively. In a software defect prediction system, it faces many difficulties in identifying the defect modules like correlation, irrelevance aspects, data redundancy, and missing samples or values. Consequently, many researchers are designed to detect software bug prediction that categorises faulty as well as non-faulty modules with software matrices. But, there are only a few works focussed to mitigate the class imbalance problem in bug prediction. In order to overcome the problem, it is required to develop an efficient software bug prediction method with the enhanced classifier. For this experimentation, the input data are taken from the standard online data sources. Initially, the input data undergo pre-processing phase and then, the pre-processed data are provided as input to the feature extraction by utilising the Auto-Encoder. These obtained features are utilised in getting the optimal fused features with the help of a new Hybrid Honey Badger Cat Swarm Algorithm (HHBCSA). Finally, these features are fed as input to the Optimised Parallel Cascaded Deep Network (OPCDP), where the “Extreme Learning Machine (ELM) and Deep Belief Network (DBN)” are used for the prediction of software bugs, in which the parameters from both classifiers are optimised by proposed HHBCSA algorithm. From the investigations, the recommended software bug prediction method offers a quicker bug prediction result, which helps to detect and remove the software bug easily and accurately.

In today’s rapid development of global technology, the global demand for scientific and technological talents remains high. To provide a reliable talent referral channel, a talent recommendation model based on Bidirectional Encoder Representation from Transformers (BERT) and Bi-directional Long Short-Term Memory (BLSTM) was constructed. This model enables the matching of talented individuals with job opportunities. The results demonstrated that the accuracy and F1 value of BLSTM-BERT in the test set were 0.95 and 0.92, respectively. The precision rate, recall rate, F1-socre value and accuracy rate of BLSTM-CNN model were 0.96, 0.97, 0.96 and 0.97, respectively. The correct prediction rate of the talent recommendation model for the four types of talents was 1.0. It is evident that the talent recommendation model has a high accuracy in predicting talent categories and can precisely recommend necessary scientific and technological professionals for businesses.

Knowledge Tracing (KT) represents an analysis of the state in terms of knowledge among students to predict if the student can answer a problem based on test results. Generally, a human teacher tracks the knowledge of students and customises the teaching based on the needs of the students. Nowadays, the rise of online education platforms leads to the development of machines for tracking the knowledge of students and improving their learning experience. The accuracy of the classical KT techniques needs to be improved. Thus, this paper implemented the Student Psychology Teaching Learning Optimisation-based Deep Long Short-Term Memory (SPTLO-based DLSTM) for predicting student performance. Here, $Z$ -score normalisation is adapted for performing normalisation of data to make the data value rely on a specific range. Furthermore, the Synthetic Minority Oversampling Technique (SMOTE) is engaged to augment data to make data apt for enhanced handling. The Deep Maxout Network (DMN) with Ruzicka similarity is considered for feature fusion. The integration of deep KT to predict student performance is executed with Deep Long Short-Term Memory (DLSTM), which is trained to employ SPTLO. The SPTLO is generated by unifying Student Psychology Based Optimisation (SPBO) and Teaching-Learning-Based Optimisation (TLBO). Here, SPTLO-based DLSTM presented supreme accuracy of 92.5%, Mean Absolute Error (MAE) of 0.064 and Root mean square error (RMSE) of 0.312.