Journal of Information & Knowledge Management

Social Networks have become an important part of people’s life as they share their day-to-day happenings, portray their opinions on various topics or find out information related to their queries. Due to the overwhelming volume of tweets generated on a daily basis, it is not possible to read all the tweets and differentiate the tweets based on the views or the attitude they portray only. The primary objective of sentiment analysis is to find out the attitude/emotion/opinion/sentiment that is present in the material provided. Commonly, the tweets can be clustered on the basis of them being positive or negative i.e. being in favour of the topic or being against the topic. The clustering and indexing of the tweets help in the organisation, searching, and summarisation of task. Twitter data are considered as Big Data and the information contained within the tweets is unstructured and if utilised properly can be very useful for educational and governance purposes. In this paper, a method is presented which clusters and then indexes the tweets on the basis of the sentiments and emoticons that are present in the tweet.

Cloud computing market is continually growing in the last years and becoming a new opportunity for business for private and public organisations. The diffusion of multi-tenants distributed systems accessible by clouds leads to the birth of some cross-organisational environments, increasing the organisation efficiency, promoting the business dynamism and reducing the costs. In spite of these advantages, this new business model drives the interest of researchers and practitioners through new critical issues. First of all, the multi-tenant distributed systems need new techniques to improve the traditional resource management distribution along the different tenants. Secondly, new approaches to the process analysis and monitoring analysed since cross-organisational environments allow various organisations to execute the same process in different variants. Hence, information about how each process variant characterised can be collected by the system and stored as process logs. The usefulness of such logs is twofold: these logs can be analysed using some process mining techniques to understand and improve the business processes and can be used to find better resource management and scalability. This paper proposes a cloud computing multi-tenancy architecture to support cross-organisational process executions and improve resource management distribution. Moreover, the approach supports the systematic extraction/composition of distributed data from the system event logs that are assumed to carry information of each process variant. To this aim, the approach also integrates an online process mining technique for the runtime extraction of business rules from event logs. Declarative processes are used to represent process variants running on the analysed infrastructure as they are particularly suited to represent the business process in a context characterised by low predictability and high variability. In this work, we also present a case study where the proposed architecture is implemented and applied to the execution of a real-life process of online products selling.

Air pollution indicates contaminated air which arises due to the effect of physical, biological or chemical alteration to the air in the atmosphere applicable both for indoors and outdoors. This situation arises when poisonous gases, dust or smoke enter into the atmosphere and make the surroundings vulnerable for any living beings as well as difficult for them to survive. Large numbers of premature deaths happen across the globe if exposed to these pollutants on a long-term basis as major portion of the cities have the pollution level above the threshold determined by World Health Organization (WHO). So appropriate measures need to be taken on a priority basis to reduce air pollution as well as save our planet. This paper proposes a novel air pollution reduction approach which collects source pollution data. After extraction of source data, it uses various databases (DBs) and then different decisions or classes are created. The decision tree was created with the help of Iterative Dichotomiser 3 (ID3) algorithm to implement the rule base appropriately depending on the air pollution level and a bunch of rule sets were derived from the decision tree further.

Water is one of the most important necessities for human survival. In municipal corporation areas, water quality affects a large part of the population. Good quality water supply is an imperative parameter that influences individuals’ health. Automated accurate water quality determination becomes an urgent necessity. Detecting the drinking water quality can prevent such scenarios prior to the critical stage. Recent research works have achieved reasonable success in predicting the water quality by deploying several machine learning-based techniques and utilising different aspects to analyse water quality. The accuracy levels of already proposed models are to be improved, keeping in mind the sensitivity of the problem domain. In the current work, Non-dominated Sorting Genetic Algorithm-II (NN-NSGA-II) was employed to train the artificial neural network (ANN) to improve its performance over its traditional counterparts. The proposed model gradually minimises two different objective functions, namely the root mean square error (RMSE) and Maximum Error (ME) in order to find the optimal weight vector for the ANN. The proposed model was compared with another two well-established models namely ANN trained with Genetic Algorithm (NN-GA) and ANN trained with Particle Swarm Optimisation (NN-PSO) in terms of accuracy, precision, recall, $F$-Measure, Matthews correlation coefficient (MCC) and Fowlkes–Mallows (FM) index. Furthermore, a data perturbation-based stability analysis is proposed to test the stability of the proposed method. The simulation results established superior accuracy of NN-NSGA-II over the other models.

Cloud computing encompasses powerful technology to perform complex computing for large applications. It provides access to various resources from any location with reduced Information Communication Technology overhead, enabling availability and scalability of resources based on consumer needs. Therefore, it has brought a paradigm shift in the way computing services are delivered. However, due to its increasing usage and rising expectations from cloud service provisioning, providing optimal and effective service is becoming an arduous task for service providers. Thus, maintaining quality of service (QoS) through regular monitoring is the utmost priority for the service provider. The key ingredient of managing QoS is a service level agreement (SLA). SLAs form a basis of measuring service quality. The cloud can be leveraged for provisioning e-Learning services to serve millions of users. The e-Learning service providers can enter into a contract with available cloud providers, which is transparent to the learners. Hence, it is critical to focus on the SLAs that need to be formulated and managed by the service providers for consumers of their e-Learning services. The objective of this paper is to propose SLA-based e-Learning service framework. This paper also identifies SLA parameters and related metrics to measure SLAs relevant to the provisioning of the proposed e-Learning service. It also proposes an innovative e-Learning service monitoring framework for managing QoS. This paper further evaluates the effectiveness of the proposed framework by provisioning of e-Learning service on Google App Engine cloud platform. It also presents experimental analysis using response graph in order to monitor associated SLAs in the proposed framework.

The surge in the use of social media tools by most businesses and corporate society for varied purposes cannot be over emphasised. The two top social media sites heavily patronised by businesses are Facebook and Twitter. For companies to harness the business potential of social media to increase competitive advantage, sentiments behind textual data of their customers, fans and competitors must be monitored and analysed with keen interest. This paper demonstrates how companies in the Telecommunication industry can understand consumer opinions, frustrations and satisfaction through opinion mining analyses and interpret customers’ textual data to enhance competitiveness. Sentiment analysis that classifies positive, negative and neutral sentiments of customers of the top three telecommunication companies in Ghana (MTN, Vodafone and Tigo) is studied. The proposed method extracts “intelligence” from the classified customers’ comments and compares it with responses from the companies. The results show how customer sentiments can be harnessed into successful online advertising projects. Companies can use the results to enhance their responsiveness to customer-centred, improve on the quality of their service, integrate social sentiments into PR plan, develop a strategy for social media marketing and leverage on the advantages of online advertising.

Since the introduction of Web 2.0, group recommendation systems become an effective tool for consulting and recommending items according to the choices of group of likeminded users. However, the population of dataset consisting of the large number of choices increases the size of storage. As a result, identification of the combination for specific recommendation becomes complex. Hence, the existing group recommendation system should support methodology for handling large data volume with varsity. In this paper, we propose a content-boosted modified termite colony optimisation-based rating prediction algorithm (CMTRP) for group recommendation system. CMTRP employs a hybrid recommendation framework with respect to the big data paradigm to deal with the trend of large data. The framework utilises the communal ratings that help to overcome the scalability problem. The experimental results reveal that CMTRP provides less error in the rating prediction and higher recommendation precision compared with the existing algorithms.

Recently internal controls, corporate governance and risk management have received a great deal of attention. Regarding internal control, several research studies address the issue of internal audit quality. Noteworthy, according to Sarbanes–Oxley (SOX) the internal controls over financial reporting are assessed by the auditors and the management. In the present study, we assess internal controls over financial reporting by employing Text Mining techniques. We analyse the annual reports of 133 publicly traded Greek Companies. The textual parts of the annual reports that refer to internal audit mechanism are extracted. We adopt a Vector Space model and the term-document matrix records the occurrence frequencies of the terms. By applying feature selection, a set of significant keywords, which are used as predictors, is extracted. The Linear Regression model developed explains the variance of the data and highlights significant predictors. The model manages to successfully assess the internal audit function. By performing PCA, major underlying procedures and concepts related to internal audit quality are revealed. Inspite of the undoubted importance of the assessment of internal audit, no previous attempt has been made to assess internal audit and to extract internal audit information from corporate disclosures by using Text Mining techniques. Our results can be useful to internal and external auditors, managers, company decision-makers, regulators and researchers.

Recently, MapReduce-based implementations of clustering algorithms have been developed to cope with the Big Data phenomenon, and they show promising results particularly for the document clustering problem. In this paper, we extend an efficient data partitioning method based on the relational analysis (RA) approach and applied to the document clustering problem, called PDC-Transitive. The introduced heuristic is parallelised using the MapReduce model iteratively and designed with a single reducer which represents a bottleneck when processing large data, we improved the design of the PDC-Transitive method to avoid the data dependencies and reduce the computation cost. Experiment results on benchmark datasets demonstrate that the enhanced heuristic yields better quality results and requires less computing time compared to the original method.

With the advent of computers and the information age, statistical and analytical problems have grown in terms of both size and complexity. Challenges in core domains of data storage, organisation and searching have evolved to the new research field called data mining. Text classification using various machine learning (ML) mechanisms encounters the difficulty of the high dimensionality of attributes vector. Therefore, a feature selection technique is very much required to discard irrelevant as well as noisy attributes from the feature set vector so that the ML algorithms can work efficiently. In this paper, rough set theory (RST)-based attribute selection methodology is applied to achieve text classification goal. A hybrid method based on RST is proposed for text documents classification. Further, the proposed method’s performance is evaluated on standard datasets.

The performance of computational grids mainly depends on the resource allocation service of a resource management system. Efficient resource allocation is essential for better resource utilisation which could be for both providers and grid users. Resource allocation includes the scheduling of gridlets to the available resources. However, the biggest challenges for grid users are to select the best resources from the available grid resources and to allocate these resources for scheduling of the gridlets. To address these issues and enhance the resource utilisation process, we propose a semantic and proximity-aware fuzzy rule-based model that improves the resource utilisation in a grid environment. The model uses fuzzy techniques with four parameters such as semantic similarity, proximity, number of total machines and number of total processors of each machine. The experimental results provide promising results. Overall, the proposed semantic and proximity-aware fuzzy rule-based decentralised resource discovery model improves the resource utilisation by 23% as compared to non-fuzzy first come first serve (FCFS) technique in a computational grid environment.