A hybrid discretization method for naïve Bayesian classifiers

Since naïve Bayesian classifiers are suitable for processing discrete attributes, many methods have been proposed for discretizing continuous ones. However, none of the previous studies apply more than one discretization method to the continuous attributes in a data set for naïve Bayesian classifiers. Different approaches employ different information embedded in continuous attributes to determine the boundaries for discretization. It is likely that discretizing the continuous attributes in a data set using different methods can utilize the information embedded in the attributes more thoroughly and thus improve the performance of naïve Bayesian classifiers. In this study, we propose a nonparametric measure to evaluate the dependence level between a continuous attribute and the class. The nonparametric measure is then used to develop a hybrid method for discretizing continuous attributes so that the accuracy of the naïve Bayesian classifier can be enhanced. This hybrid method is tested on 20 data sets, and the results demonstrate that discretizing the continuous attributes in a data set by various methods can generally have a higher prediction accuracy.     

The stochastic approximation method for adaptive Bayesian classifiers: towards online brain�Ccomputer interfaces

Recent developments of brain–computer interfaces (BCIs) bring forward some challenging problems to the machine learning community, of which classification of time-varying electrophysiological signals is a crucial one. Constructing adaptive classifiers is a promising approach to deal with this problem. In this paper, Bayesian classifiers with Gaussian mixture models (GMMs) are adopted to classify electroencephalogram (EEG) signals online. We propose to use the stochastic approximation method (SAM) as the specific gradient descent method for parameter update and systematically derive the instantaneous gradient formulas with respect to mean values and covariance matrices in the distributions of a GMM. With SAM, the parameters of mean values and covariance matrices embodied in the Bayesian classifiers can be simultaneously updated in a batch mode. The online simulation of EEG classification tasks in a BCI shows the effectiveness of the proposed SAM.     

Corporate website accessibility: does legislation matter?

Over 600 million people worldwide have disabilities ranging from visual and hearing impairments to cognitive and motor skill issues. This number is only growing as “Baby Boomers” age. Previous research reveals that those organizations, such as federal agencies and colleges, which are mandated to have accessible websites, do indeed have higher levels of accessible websites than corporate sites do. This led to the current research, which aims at understanding what factors truly impact a company’s decision to provide an accessible website. The results of a global survey of managers from a variety of industries uncovered that the key factors for influencing a company’s level of website accessibility are the number of IT professionals employed by the firm, the level of accessibility testing performed, and whether the company is mandated to have an accessible website.     

Capturing the participation and social dimensions of computer-supported collaborative learning through social network analysis: which method and measures matter?

International Journal of Computer-Supported Collaborative Learning - The increasing use of digital learning tools and platforms in formal and informal learning settings has provided broad access to...     

A framework for monitoring classifiers’ performance: when and why failure occurs?

Classifier error is the product of model bias and data variance. While understanding the bias involved when selecting a given learning algorithm, it is similarly important to understand the variability in data over time, since even the One True Model might perform poorly when training and evaluation samples diverge. Thus, it becomes the ability to identify distributional divergence is critical towards pinpointing when fracture points in classifier performance will occur, particularly since contemporary methods such as tenfolds and hold-out are poor predictors in divergent circumstances. This article implement a comprehensive evaluation framework to proactively detect breakpoints in classifiers’ predictions and shifts in data distributions through a series of statistical tests. We outline and utilize three scenarios under which data changes: sample selection bias, covariate shift, and shifting class priors. We evaluate the framework with a variety of classifiers and datasets.

---

User experience: does it matter in complex systems?

This paper explores the concept of user experience (UX) within complex systems in the domain of nuclear power production. The specific context is the development of main control room solutions. As UX has become a prevalent concept in human–computer interaction and related research fields, it is a valid question, whether the concept could bring something new to the research concerning complex safety–critical systems also. In this paper, we introduce the concept of UX as an indicator of the users’ subjective feeling of the appropriateness of the proposed tool for the activity. Based on activity theoretical contemplation, we presume that UX is grounded in the feelings and emotions evoked in the usage of the systems and which concern the status of the whole activity. In the paper, a theory-based characterization of UXs in the particular domain is developed based on three general functions of a tool: instrument, psychological, and communicative. We present the operationalization of UX and three studies in which UXs concerning different control room systems during a control room transformation process were followed. Based on the results of the three studies, we find that the significance of UX as an indicator of quality in use lies in the ability to bring the professional users’ experiences, which are embedded in the inner characteristics of the work and not always observable by external evaluators, to the process of designing new systems. Evidence to support the initial background assumption that UXs may concern instrumental aspects of tools was identified in the studies.     

Force-endurance relationship: does it matter if gloves are donned?

The human hand is a very useful multipurpose tool in all environments. However, performance capabilities are compromised considerably when gloves are donned. This is especially true for extra-vehicular activity (EVA) gloves used in a space environment. The primary aim of this study was to establish exertion and endurance limits for specific tasks. The objective of this study was to develop grip force endurance relations. Six subjects participated in a factorial experiment involving three hand conditions, three pressure differentials, and four levels of force exertion. The results indicate that while the force that could be exerted depended on the glove, pressure differential, and the level of exertion, the endurance time at any exertion level depended just on the level of exertion expressed as a percentage of maximum exertion possible at that condition. The impact of these findings for practitioners as well as theoreticians is discussed.     

Ergodicity of spike trains: when does trial averaging make sense?

Neuronal information processing is often studied on the basis of spiking patterns. The relevant statistics such as firing rates calculated with the peri-stimulus time histogram are obtained by averaging spiking patterns over many experimental runs. However, animals should respond to one experimental stimulation in real situations, and what is available to the brain is not the trial statistics but the population statistics. Consequently, physiological ergodicity, namely, the consistency between trial averaging and population averaging, is implicitly assumed in the data analyses, although it does not trivially hold true. In this letter, we investigate how characteristics of noisy neural network models, such as single neuron properties, external stimuli, and synaptic inputs, affect the statistics of firing patterns. In particular, we show that how high membrane potential sensitivity to input fluctuations, inability of neurons to remember past inputs, external stimuli with large variability and temporally separated peaks, and relatively few contributions of synaptic inputs result in spike trains that are reproducible over many trials. The reproducibility of spike trains and synchronous firing are contrasted and related to the ergodicity issue. Several numerical calculations with neural network examples are carried out to support the theoretical results.     

How does organisational absorptive capacity matter in the assimilation of enterprise information systems?

Extant literature offers two mostly distinct perspectives on enterprise systems assimilation – driven either by internal expertise and learning capability or by external institutional pressures. This study combines the two perspectives and subscribes to the view that organisations’ learning capability moderates their acquiescence to institutional pressures. The study then anchors organisational learning capability to the concept of absorptive capacity and proposes that its two dimensions – potential absorptive capacity (PACAP) and realised absorptive capacity (RACAP) – affect enterprise systems assimilation through different pathways. Our survey‐based empirical study of Enterprise Resource Planning (ERP) systems in the post‐implementation stage reveals that while both PACAP and RACAP have a positive direct impact on assimilation, PACAP positively moderates the impact of mimetic (institutional) pressures, but not normative (institutional) pressures, on assimilation; whereas RACAP positively moderates the impact of normative pressures, but not mimetic pressures, on assimilation. Thus, our theoretical contribution lies in understanding the distinct ways in which PACAP and RACAP moderate the influence of external institutional pressures on enterprise systems assimilation.     

Gender differences in computer attitudes: Does the school matter?

In most western countries, the participation of females in ICT professional careers is not only low but is also still falling [Anderson, N., Lankhear, C., Timms, C., & Courtney, L. (in press). Because it’s boring, irrelevant and I don’t like computers’: Why high school girls avoid professionally-oriented ICT subjects. Computers & Education.]. Policy makers as well as researchers often assume that the interest of girls in computing and ICT-professions could be increased at school. For example, female teachers who are confident ICT-users, are expected to act as positive role models for girls. However, because most of the research on gender and computing has been focussing on the influence of none-school related factors, there is little empirical evidence that schools or teachers are able to influence girls’ attitude toward ICT. Using the data of a Dutch large-scale survey on ICT use in primary education (almost 4000 grade 5 students), this study explores the influence of both none-school related factors and school related factors on students’ computer attitude. Although the between-school variance of girls’ computer attitude is higher than that of boys’ computer attitude, multilevel analyses show that most of the variance in computer attitude is explained by none-school related student factors. Two school related factors turned out to have a small positive effect on the computer attitude of girls: a teacher-centred pedagogical approach and the computer experience of female teachers.     

Sexual self-presentation on social network sites: Who does it and how is it perceived?

The aims of this study were to investigate (a) the role of peer factors in adolescents’ sexual self-presentation on social network sites, and (b) how adolescents who present themselves sexually online are evaluated by others. 238 high school students (aged 12–18) evaluated either a sexual or non-sexual online presentation of same-sex and opposite-sex peers. Moreover, they filled in a questionnaire about their own self-presentation behavior. Findings showed that peer norms played a crucial role in whether adolescents posted sexual pictures of themselves online. Moreover, need for popularity was a strong predictor for posting such pictures. Girls who presented themselves in sexual ways were evaluated more negatively by other girls but more positively by boys. Similarly, boys who presented themselves in sexual ways were evaluated more positively by girls but not by boys.     

Visual quality assessment algorithms: what does the future hold?

Creating algorithms capable of predicting the perceived quality of a visual stimulus defines the field of objective visual quality assessment (QA). The field of objective QA has received tremendous attention in the recent past, with many successful algorithms being proposed for this purpose. Our concern here is not with the past however; in this paper we discuss our vision for the future of visual quality assessment research. We first introduce the area of quality assessment and state its relevance. We describe current standards for gauging algorithmic performance and define terms that we will use through this paper. We then journey through 2D image and video quality assessment. We summarize recent approaches to these problems and discuss in detail our vision for future research on the problems of full-reference and no-reference 2D image and video quality assessment. From there, we move on to the currently popular area of 3D QA. We discuss recent databases, algorithms and 3D quality of experience. This yet-nascent technology provides for tremendous scope in terms of research activities and we summarize each of them. We then move on to more esoteric topics such as algorithmic assessment of aesthetics in natural images and in art. We discuss current research and hypothesize about possible paths to tread. Towards the end of this article, we discuss some other areas of interest including high-definition (HD) quality assessment, immersive environments and so on before summarizing interesting avenues for future work in multimedia (i.e., audio-visual) quality assessment.     

Component meets service: what does the mongrel look like?

Computing with services has attracted much attention as a promising approach for developing distributed applications. The approach is often advertised as being superior to distributed component-based software engineering (CBSE), because it provides a higher potential to bridge heterogeneous IT application and infrastructure landscapes. It facilitates cross-institutional cooperation, lets services run over all kinds of ubiquitous communication infrastructure, scales better and simplifies legacy software integration. If this were absolutely true, there would be no reason for a consortium of major vendors of service and Java EE technology to come up with a new specification, called service component architecture (SCA). This emerging standard tries to leverage service-oriented architecture (SOA) principles with component-based software development techniques. In this article we discuss some commonalities and fundamental differences of the CBS and SOA worlds. We illustrate SCA briefly using snippets of an ongoing case study based on an e-university federation. Then we elaborate on the qualities and current deficits of SCA in the light of CBSE findings and related works.     

Having expectations of information systems benefits that match received benefits: does it really matter?

A study was conducted to examine the effect of implementing a new system on its users, specifically, the relationship between pre-implementation expectations and their perceived benefits based on post-implementation experience. Disconfirmation theory was used as the theoretical basis; this predicts that unrealistically high expectations will result in lower levels of perceived benefit than those associated with realistic expectations (i.e. where expectations match experience). Support was found for this prediction, refuting the predictions of dissonance theory. In addition to examining expectations of system use generally, six expectation categories were examined to identify the critical categories where managers should keep expectations from becoming unrealistically high. Significant relationships were found for three expectation categories: system usefulness, ease of use, and information quality. The results indicate that creating and maintaining realistic expectations of future system benefits really does matter.     

Setting the stage for the metacognition during hypermedia learning: What motivation constructs matter?

Think-aloud and self-report data from 85 undergraduates were used to examine the relationship between motivation constructs and metacognition during hypermedia learning. Participants used hypermedia for 30 min to learn about the circulatory system. Think-aloud data were collected during this 30-min learning task to determine the extent to which participants used metacognitive processes related to monitoring: their understanding, the environment, and goals. Additionally, participants completed a self-report questionnaire, which measured various motivation constructs. Results from stepwise regressions indicated that self-efficacy significantly predicted the extent to which participants monitored emerging understanding and relevancy of content in the environment. Additionally, results indicated that extrinsic motivation significantly predicted the extent to which participants monitored their learning task goals with hypermedia. Lastly, results indicated a significant, positive relationship between self-efficacy and prior domain knowledge.     

RETRACTED ARTICLE: Biometric identification,financial inclusion and economic growth in India: does mobile penetration matter?

ABSTRACT

Using data on major Indian states during 2001–2014, the paper analyzes the impact of AIM (Aadhaar, Inclusion and Mobile) trinity on economic growth. Using advanced panel data techniques, the evidence suggests that mobile penetration exerts a positive and statistically significant impact on economic growth. In addition, there exist strong complementarities between mobile telephony and financial inclusion. The magnitude of the complementarity is significantly higher in case of financial use as compared with financial access. In addition, the evidence points to the fact that biometric identification process (Aadhaar) is making its impact felt on financial inclusion, primarily by improving financial access.     

Using pencil and paper,Internet and touch-tone phones for self-administered surveys: does methodology matter?

To compare responses to personally sensitive questions, 352 undergraduates were randomly assigned to respond anonymously to a survey using one of three survey methods: pencil and paper mail-in, Internet survey, or an automated touch-tone telephone response system. The survey contained 68 brief Yes/No questions ranging from low to high sensitivity in 13 domains, such as general honesty, academic honesty, prejudice, illegal behavior, alcohol use, substance use, violence, and sexual behavior. We found no significant differences (p>0.05) in participants’ responses among these three media for any of the questions. This suggests that for some populations, under some circumstances, Internet and touch-tone telephone systems achieve the same results as traditional pencil and paper surveys.