Search

Emotion detection and Sentiment analysis techniques are used to understand polarity or emotions expressed by people in many cases, especially during interactive systems use. Recognizing users' emotions is an important topic for human–computer interaction. Computers that recognize emotions would provide more natural interactions. Also, emotion detection helps design human-centred systems that provide adaptable behaviour change interventions based on users' emotions. The growing capability of machine learning to analyze big data and extract emotions therein has led to a surge in research in this domain. With this increased attention, it becomes essential to investigate this research area and provide a comprehensive review of the current state. In this paper, we conduct a systematic review of 123 papers on machine learning-based emotion detection to investigate research trends along many themes, including machine learning approaches, application domain, data, evaluation, and outcome. The results demonstrate: 1) increasing interest in this domain, 2) supervised machine learning (namely, SVM and Naïve Bayes) are the most popular algorithms, 3) Text datasets in the English language are the most common data source, and 4) most research use Accuracy to evaluate performance. Based on the findings, we suggest future directions and recommendations for developing human-centred systems. [ABSTRACT FROM AUTHOR]

This paper explores the different insider employee-led cyber frauds (IECF) based on the recent large-scale fraud events of prominent Indian banking institutions. Examining the different types of fraud and appropriate control measures will protect the banking industry from fraudsters. In this study, we identify and classify Cyber Fraud (CF), map the severity of the fraud on a scale of priority, test the mitigation effectiveness, and propose optimal mitigation measures. The identification and classification of CF losses were based on a literature review and focus group discussions with risk and vigilance officers and cyber cell experts. The CF was analyzed using secondary data. We predicted and prioritized CF based on machine learning-derived Random Forest (RF). An efficient fraud mitigation model was developed based on an offender-victim-centric approach. Mitigation is advised both before and after fraud occurs. Through the findings of this research, banks and fraud investigators can prevent CF by detecting it quickly and controlling it on time. This study proposes a structured, sustainable CF mitigation plan that protects banks, employees, regulators, customers, and the economy, thus saving time, resources, and money. Further, these mitigation measures will improve the reputation of the Indian banking industry and ensure its survival. [ABSTRACT FROM AUTHOR]

With the rapid development of social platforms and human activities, people are now interested to solve their problems with the help of the social crowd-powered system. So, crowdsourcing has become a promising way of solving problems in a distributed manner within a specific time. Crowd workers pick up a task (simple or complex) and solve it with competing interest or collaboratively and receive an incentive (as monetary or non-monetary). In paid crowdsourcing platforms, crowd workers solve complex tasks and get remunerations through prior bidding or prior announcement of fee decided by the requester (task provider). For decomposable tasks, a single winner may not provide a significant solution to the requester due to the insufficient knowledge. So, we induce collaboration in competitive crowdsourcing markets to better handle decomposable-type tasks. In this paper, we propose an envelope game-based mechanism that ensures if the tasks are decomposable then the workers will be encouraged to collaborate with their subtasks and share their remunerations. This mechanism also increases the chance of receiving more number of cost-effective solutions. Thus both the requester and crowd workers get benefits from the system. The effectiveness of the proposed mechanism is evaluated through empirical analysis on simulated environments. [ABSTRACT FROM AUTHOR]

This paper examines how the use of multiple platforms is tied to awareness of algorithms. It builds on the premise that users interact with ecologies or environments of technologies rather than single platforms. The study also supplements work on algorithmic awareness by implementing a mixed-method study to account for how Costa Rican users of Netflix and Spotify understood and related to the algorithms of these platforms. This study combined a survey of 258 participants and 21 in-depth semistructured interviews. Findings demonstrate that multi-platform users were more aware of algorithms and carried out more practical actions to obtain algorithmic recommendations than single-platform users. Although user type did not predict participants' attitudes towards algorithmic recommendations, higher levels of awareness were associated with more positive attitudes towards algorithms. The study also shows that differences in levels of awareness explained users' emotional arousal derived from algorithms. [ABSTRACT FROM AUTHOR]

Research has proven that having high level of emotional intelligence (EI) can reduce the chance of getting mental illness. EI, and its component, can be improved with training, but currently the process is less flexible and very time-consuming. Machine learning (ML), on the other hand, can analyse huge amount of data to discover useful trends and patterns in shortest time possible. Despite the benefits, ML usage in EI training is scarce. In this paper, we studied 92 journal articles to discover the trend of the ML utilisation in the study of EI and its components. This survey aims to pave way for future studies that could lead to implementation of ML in EI training, and to rope in researchers in psychology and computer science to find possibilities of having a generic ML algorithm for every EI's components. Our findings show an increasing trend to apply ML on EI components, and Support Vector Machine and Neural Network are the two most popular ML algorithms used in those researches. We also found that social skill and empathy are the least exposed EI components to ML. Finally, we provide recommendations for future research direction of ML in EI domain, and EI in ML. [ABSTRACT FROM AUTHOR]

In this paper, we propose a method for user interface (UI) behaviour prediction in commercial applications. The proposed method predicts appropriate UI behaviours for an application by learning repeated UI behaviour sequences from users. To this end, we adopted the long short-term memory algorithm based on the evaluation of a keystroke-level model. Our prediction model takes up to seven consecutive actions as inputs to predict the final UI actions that a user is likely to perform. We verified the effectiveness of the proposed method for both PC applications and mobile game environments. Our experimental results demonstrate that the proposed system can predict user UI behaviours in an application on the client side and provide useful behavioural information for optimising UI layouts. [ABSTRACT FROM AUTHOR]

This article focuses on the challenging topic of indoor location and guidance in daily tasks of location and hostile scenarios (i.e. emergency evacuations) that could be produced, which at present have no universally viable and implemented solution. Indoor guidance in public spaces such as hospitals or private spaces such as a company building helps people with mild and moderate cognitive disabilities to get on safely in their daily lives and to promote their social integration. This paper analyses three areas of research: location technologies, location services for sensory navigation and problems derived from stress due to panic in hostile circumstances. Therefore, in this work we focus on finding solutions related to the absence of GNSS signals, discriminating of different floors in a building, problems of adequate information transmission to the users and absence of connectivity and interfaces, i.e. we must adapt the solutions that exist for outdoor to indoor. In particular, the analysis has focused on viable technologies for emergency or everyday situations that create a network of roads within a single building. The results obtained outline the requirements that need to be addressed in the development of wayfinding services that provides sensory guidance to people with cognitive disabilities. [ABSTRACT FROM AUTHOR]

The ability to exploit students' sentiments using different machine learning techniques is considered an important strategy for planning and manoeuvring in a collaborative educational environment. The advancement of machine learning technology is energised by the healthy growth of big data technologies. This helps the applications based on Sentiment Mining (SM) using big data to become a common platform for data mining activities. However, very little has been studied on the sentiment application using a huge amount of available educational data. Therefore, this paper has made an attempt to mine the academic data using different efficient machine learning algorithms. The contribution of this paper is two-fold: (i) studying the sentiment polarity (positive, negative and neutral) from students' data using machine learning techniques, and (ii) modelling and predicting students' emotions (Amused, Anxiety, Bored, Confused, Enthused, Excited, Frustrated, etc.) using the big data frameworks. The developed SM techniques using big data frameworks can be scaled and made adaptable for source variation, velocity and veracity to maximise value mining for the benefit of students, faculties and other stakeholders. [ABSTRACT FROM AUTHOR]

There are great interests of designing research metrics and indices to measure the research impacts in research institutes. Unfortunately, most of those indices ignore critical design issues, e.g. the disparity between domains, the impact of journals or conferences in which papers are published, normalising the range of the index values to certain intervals, and the scalability of using the index to rank different research entities. In this paper, a new normalised fuzzy index, (NFindex), is proposed as a fuzzy-based research impact metric. The proposed index is a scalable index whose values are normalised to the percentage levels. NFindex achieves both inter-discipline normalisation and intra-discipline consistency. The capability of NFindex to achieve the inter-discipline normalisation enables fair comparison between different research domains regardless their nature in terms of influence and contribution to other research areas, e.g. natural science. Therefore, NFindex gives a universal normalised single-number metric that can be used by research institutes to solve the problem of inter-discipline scholar ranking. Moreover, it can help universal ranking of universities and research institutes according to their research capabilities and impacts. The obtained results, on diverse research areas, prove the potential of NFindex in terms of both intra-discipline consistency and inter-discipline normalisation. [ABSTRACT FROM AUTHOR]

The research field of crisis informatics examines, amongst others, the potentials and barriers of social media use during conflicts and crises. Social media allow emergency services to reach the public easily in the context of crisis communication and receive valuable information (e.g. pictures) from social media data. However, the vast amount of data generated during large-scale incidents can lead to issues of information overload and quality. To mitigate these issues, this paper proposes the semi-automatic creation of alerts including keyword, relevance and information quality filters based on cross-platform social media data. We conducted empirical studies and workshops with emergency services across Europe to raise requirements, then iteratively designed and implemented an approach to support emergency services, and performed multiple evaluations, including live demonstrations and field trials, to research the potentials of social media-based alerts. Finally, we present the findings and implications based on semi-structured interviews with emergency services, highlighting the need for usable configurability and white-box algorithm representation. [ABSTRACT FROM AUTHOR]

A key agenda of policy-makers in both the developing and the developed worlds is to have an inclusive growth. Financial services have been recognised as one of the fundamental services that have the potential to achieve this objective and help move towards a more just and equitable society. This paper attempts to assess the impact of three information and communication technology-driven initiatives in the field of financial services by differentiating them on the basis of service design. This study goes beyond the stated objective of these services and uses Sen’s capability perspective to study their impact across three non-economic outcome variables viz. literacy, service-specific well-being and empowerment. The data for the study are collected from the urban poor across different locations in a large metropolitan city in South India where the three different services are being offered. The study follows a quasi-experimental design (field study) and ensures matching of the treatment and control groups through propensity score matching. The matched data are analysed using an independent samplet-test. The result of the analysis shows that service design plays a key role in increasing service literacy, improving service-specific well-being and empowering end users. The study suggests that the true challenge for providers (government and partner agencies) of these services is in designing a service that ensures a balance of structure and flexibility without constraining the choice of end users through innovative service blueprints. [ABSTRACT FROM AUTHOR]

The almost unlimited access to educational information plethora came with a drawback: finding meaningful material is not a straightforward task anymore. Recommender algorithms can be used to make smart decisions in complex information systems and help the users decide upon useful materials; therefore, they become a promising area in academia and industry. The current paper presents a survey on educational recommender systems (RS): a set of analysis criteria are exposed and the technological specifications and challenges of each analysed system are provided, in the context of the main trends in the development of RS. Also, an ontology-based educational recommendation mechanism is proposed and its application to lifelong learning is highlighted, proving that RS can successfully support new learning paradigms. [ABSTRACT FROM PUBLISHER]

Hijacked journals are cloned websites that resemble the homepages of legitimate journals, whose aim is to collect processing and publication fees from unwary authors. There is a growing recognition that the recent proliferation of these scam sites poses a threat to the integrity of the scientific process. This study presents an approach intended to prevent the indexing of papers published by hijacked journals in scientific databases by using classification algorithms. We will provide an overview of the problem, define key features of hijacked journals, and present a decision tree that can be used to detect hijacked publications. [ABSTRACT FROM PUBLISHER]

Emotion is a cognitive process and is one of the important characteristics of human beings that makes them different from machines. Traditionally, interactions between humans and machines like computers do not exhibit any emotional exchanges. If we could build any system that is intelligent enough to interact with humans that involves emotions, that is, it can detect user emotions and change its behaviour accordingly, then using machines could be more effective and friendly. Many approaches have been taken to detect user emotions. Affective computing is the field that detects user emotion in a particular moment. Our approach in this paper is to detect user emotions by analysing the keyboard typing patterns of the user and the type of texts (words, sentences) typed by them. This combined analysis gives us a promising result showing a substantial number of emotional states detected from user input. Several machine learning algorithms were used to analyse keystroke timing attributes and text pattern. We have chosen keystroke because it is the cheapest and most available medium to interact with computers. We have considered seven emotional classes for classifying the emotional states. For text pattern analysis, we have used vector space model with Jaccard similarity method to classify free-text input. Our combined approach showed above 80% accuracies in identifying emotions. [ABSTRACT FROM PUBLISHER]

The perception of the visual complexity of World Wide Web (Web) pages is a topic of significant interest. Previous work has examined the relationship between complexity and various aspects of presentation, including font styles, colours and images, but automatically quantifying this dimension of a web page at the level of the document remains a challenge. In this paper we demonstrate that areas of high complexity can be identified by detecting areas, or ‘chunks’, of a web page high in block-level elements. We report a computational algorithm that captures this metric and places web pages in a sequence that shows an 86% correlation with the sequences generated through user judgements of complexity. The work shows that structural aspects of a web page influence how complex a user perceives it to be, and presents a straightforward means of determining complexity through examining the DOM. [ABSTRACT FROM PUBLISHER]

Mining sequential patterns is to discover sequential purchasing behaviours for most of the customers from a large number of customer transactions. The strategy of mining sequential patterns focuses on discovering frequent sequences. A frequent sequence is an ordered list of the itemsets purchased by a sufficient number of customers. The previous approaches for mining sequential patterns need to repeatedly scan the database so that they take a large amount of computation time to find frequent sequences. The customer transactions will grow rapidly in a short time, and some of the customer transactions may be antiquated. Consequently, the frequent sequences may be changed due to the insertion of new customer transactions or the deletion of old customer transactions from the database. It may require rediscovering all the patterns by scanning the entire updated customer transaction database. In this paper, we propose an incremental updating technique to maintain the discovered sequential patterns when transactions are inserted into or deleted from the database. Our approach partitions the database into some segments and scans the database segment by segment. For each segment scan, our approach prunes those sequences that cannot be frequent sequences any more to accelerate the finding process of the frequent sequences. Therefore, the number of database scans can be significantly reduced by our approach. The experimental results show that our algorithms are more efficient than other algorithms for the maintenance of mining sequential patterns. [ABSTRACT FROM AUTHOR]