Political polarisation has become an increasingly alarming issue in society, exacerbated by the widespread use of social media and the development of filter bubbles among social media users. This environment has left users susceptible to disinformation, especially those with whom a user is politically aligned. In this research, we integrate truth bias, elaboration likelihood model and new media literacy into a model for explaining social media engagement (with both disinformation and factual information) and analysing how political polarisation (operationalised as political alignment between users) influences perceptions and behaviours. Using an experimental design, we analyse the model separately for posts containing disinformation and factual information, highlighting key differences. Political alignment positively moderates truth bias's effect on engagement with disinformation. For both disinformation and factual information, political alignment moderates the effect of generalised communicative suspicion (GCS) on truth bias, such that GCS's effect on truth bias flips from negative to positive as political alignment increases. Issue involvement and political alignment appear to be the primary drivers of disinformation engagement, with critical consuming media literacy failing to mitigate engagement. Our findings contribute to the understanding of persuasion, conviction, amplification, polarisation and aversion related to fake news on social media.

Due to the nature of modern software engineering, automated techniques for detecting software vulnerabilities are necessary for developing secure systems. While deep learning approaches have been designed to address these issues, many focus solely on source and binary versions of code, ignoring intermediate representations. Similarly, models exist that evaluate images created from code, yet they fail to provide multiclass classification of vulnerabilities, which is necessary for allowing developers to address specific insecurities. This research seeks to fill this gap by determining the accuracy of deep learning approaches, performing binary and multiclass classification of images generated from tokenized source code. To accomplish this, we performed a model-based formulative analysis of several models, comparing their accuracy using a PHP-based dataset of web-based vulnerabilities to suggest an optimal model for vulnerability detection. The research resulted in a process for creating images from PHP tokens and a ConvNext convolutional neural network that operated on 'extended' grayscale images of tokenized PHP source code. Our model achieved macro Fl scores of 0.9649 and 0.9494 in binary and multiclass classification, respectively; this approach outperformed existing models operating on the tokenized code of this same dataset. Ultimately, these results provide significant insight into novel approaches for future vulnerability detection.

Disinformation researchers have much to learn about the psychological factors that lead to users’ biases. The gap in users’ new media literacy, along with personality traits, are factors that may contribute to bias. We conducted an experiment that varied levels of articles' linguistic veracity to determine how these factors influence users' engagement behavior. We found that users with higher literacy in new media were less likely to engage with disinformation featuring low veracity language. However, users with higher literacy in traditional media were more likely to engage with disinformation, regardless of the article's linguistic veracity.

BACKGROUND:Persuasive technology is an umbrella term that encompasses software (eg, mobile apps) or hardware (eg, smartwatches) designed to influence users to perform preferable behavior once or on a long-term basis. Considering the ubiquitous nature of mobile devices across all socioeconomic groups, user behavior modification thrives under the personalized care that persuasive technology can offer. However, there is no guidance for developing personalized persuasive technologies based on the psychological characteristics of users.OBJECTIVE:This study examined the role that psychological characteristics play in interpreted mobile health (mHealth) screen perceived persuasiveness. In addition, this study aims to explore how users' psychological characteristics drive the perceived persuasiveness of digital health technologies in an effort to assist developers and researchers of digital health technologies by creating more engaging solutions.METHODS:An experiment was designed to evaluate how psychological characteristics (self-efficacy, health consciousness, health motivation, and the Big Five personality traits) affect the perceived persuasiveness of digital health technologies, using the persuasive system design framework. Participants (n=262) were recruited by Qualtrics International, Inc, using the web-based survey system of the XM Research Service. This experiment involved a survey-based design with a series of 25 mHealth app screens that featured the use of persuasive principles, with a focus on physical activity. Exploratory factor analysis and linear regression were used to evaluate the multifaceted needs of digital health users based on their psychological characteristics.RESULTS:The results imply that an individual user's psychological characteristics (self-efficacy, health consciousness, health motivation, and extraversion) affect interpreted mHealth screen perceived persuasiveness, and combinations of persuasive principles and psychological characteristics lead to greater perceived persuasiveness. The F test (ie, ANOVA) for model 1 was significant (F9,6540=191.806; P<.001), with an adjusted R2 of 0.208, indicating that the demographic variables explained 20.8% of the variance in perceived persuasiveness. Gender was a significant predictor, with women having higher perceived persuasiveness (P=.008) relative to men. Age was a significant predictor of perceived persuasiveness with individuals aged 40 to 59 years (P<.001) and ≥60 years (P<.001). Model 2 was significant (F13,6536=341.035; P<.001), with an adjusted R2 of 0.403, indicating that the demographic variables self-efficacy, health consciousness, health motivation, and extraversion together explained 40.3% of the variance in perceived persuasiveness.CONCLUSIONS:This study evaluates the role that psychological characteristics play in interpreted mHealth screen perceived persuasiveness. Findings indicate that self-efficacy, health consciousness, health motivation, extraversion, gender, age, and education significantly influence the perceived persuasiveness of digital health technologies. Moreover, this study showed that varying combinations of psychological characteristics and demographic variables affected the perceived persuasiveness of the primary persuasive technology category.

The Common Data Elements (CDEs) standard of the International organization for Standardization (ISO) 11179 is commonly used in the field of clinical data processing. The Biomedical Research Integrated Domain Group (BRIDG) model is the framework for biomedical and clinical research. Mapping CDEs to BRIDG (also known as CDE classification) would help with interoperability and data analysis in the field of clinical research. That said, manually mapping CDEs to their corresponding BRIDG class is highly time-consuming and labor-intensive. In this paper we present a new classification algorithm along with a new oversampling method. Our algorithm uses the Term Frequency-Inverse Document Frequency (TF-IDF) as the feature representation method. By assigning different weights to various attributes, we enable more important attributes to perform more important roles during the mapping process. In addition, the oversampling method generates every new attribute in the minor class by picking the length and setting the word of the new attribute according to the existing training set. Our research outcomes demonstrate significant contributions to the field in the following ways: (1) Generation of a new CDE classification algorithm that outperforms existing algorithms in the literature, including the Random Forest Classifier, Linear Support Vector Classification (SVC), Multinomial Naive Bayes (NB), Logistic Regression, and Long Short-Term Memory (LSTM) networks, in terms of accuracy, precision, recall, and F-1 score measures. (2) Generation of a new oversampling method able to improve CDE classification accuracy for Random Forest and Multinomial NB. (3) Our classification algorithm employs two novel attributes, namely “Data Element Preferred Definition” and “Document,” which are more efficient at classifying CDEs than the six attributes traditionally selected by domain experts.

It is well known that microRNAs (miRNAs or miRs) are small (~18-25 nt) yet highly potent non-coding RNA-derived RNAs (ndRNAs), originating from pre-miRNA fragmentation, that have been shown to alter the post-transcriptional functionality of many messenger RNAs (mRNAs). Biologically, the identification and study of miRNAs is very critical due to their increasing significance as biomarkers for many types of cancers and other genetic diseases. While empirical evidence supporting the existence of several novel ndRNAs excised from other longer non coding RNAs (ncRNAs) is growing, recent evidence suggests the full extent of their prevalence is likely underappreciated. Although some computational methods have been designed to help domain experts identify and understand miRNAs by analyzing Next Generation Sequencing (NGS) datasets, there are some crucial challenges, such as efficiency, effectiveness, and generalizability, in the state-of-the-art in-silico methods. To address such problems, our group proposed a new algorithm to mine ndRNAs by applying wavelet-based signal processing techniques as opposed to the current string-based NGS sequence alignment/analysis. However, due to novelty of the approach, our initial version of the algorithm was focused specifically on mining miRNAs, snoRNA-derived RNAs (sdRNAs) and transfer RNA (tRNA) fragments (tRFs) because of their importance in the literature plus the availability of experimentally validated databases to confirm our findings. Despite the computational issues, we still lack a basic understanding of the existence and the range of ndRNA functionalities from a) ndRNAs other than miRs, sdRNAs & tRFs in humans, and b) all ndRNAs in millions of organisms other than humans. Hence, there is an urgent requirement to automate the extraction and experimentation of ndRNAs, especially considering the rate at which NGS data is being produced. Therefore, in the current article, we extended our algorithm to be applicable to ~500 organisms—including eukaryotes, plants, bacteria, fungi, and protists—along with all their ncRNAs available in the current NCBI annotation. We also constructed a real-time user-friendly platform, SURFR, available at salts.soc.southalabama.edu/surfr, to aid domain experts and the aspiring biomedical scientists to perform RNA-Seq experiments to study ndRNAs. Not only our platform is extremely efficient, but we are also capable of allowing the users to identify, analyze, visualize, and compare ndRNAs from up to 30 NGS files to perform rigorous experimentation. Moreover, access to NGS files from public databases like SRA, and ndRNAs from private databases like TCGA are made readily available to the users to further validate their novel findings. Finally, we provide theoretical validation to examine our platform’s effectiveness.

Developments in virtual containers, especially in the cloud infrastructure, have led to diversification of jobs that containers are being used to support, particularly in the big data and machine learning spaces. The diversification has been powered by the adoption of orchestration systems that marshal fleets of containers to accomplish complex programming tasks. The additional components in the vertical technology stack, plus the continued horizontal scaling have led to questions regarding how to forensically analyze complicated technology stacks. This paper proposed a solution through the use of introspection. An exploratory case study has been conducted on a bare-metal cloud that utilizes Kubernetes, the introspection tool Prometheus, and Apache Spark. The contribution of this research is two-fold. First, it provides empirical support that introspection tools can acquire forensically viable data from different levels of a technology stack. Second, it provides the ground work for comparisons between different virtual container platforms.

Flexible fine-grained weather forecasting is a problem of national importance due to its stark impacts on economic development and human livelihoods. It remains challenging for such forecasting, given the limitation of currently employed statistical models, that usually involve the complex simulation governed by atmosphere physical equations. To address such a challenge, we develop a deep learning-based prediction model, called Micro-Macro, aiming to precisely forecast weather conditions in the fine temporal resolution (i.e., multiple consecutive short time horizons) based on both the atmospheric numerical output of WRF-HRRR (the weather research and forecasting model with high-resolution rapid refresh) and the ground observation of Mesonet stations. It includes: 1) an Encoder which leverages a set of LSTM units to process the past measurements sequentially in the temporal domain, arriving at a final dense vector that can capture the sequential temporal patterns; 2) a Periodical Mapper which is designed to extract the periodical patterns from past measurements; and 3) a Decoder which employs multiple LSTM units sequentially to forecast a set of weather parameters in the next few short time horizons. Our solution permits temporal scaling in weather parameter predictions flexibly, yielding precise weather forecasting in desirable temporal resolutions. It resorts to a number of Micro-Macro model instances, called modelets, one for each weather parameter per Mesonet station site, to collectively predict a target region precisely. Extensive experiments are conducted to forecast four important weather parameters at two Mesonet station sites. The results exhibit that our Micro-Macro model can achieve high prediction accuracy, outperforming almost all compared counterparts on four parameters of interest.

Action rule mining seeks to generate rules that indicate what changes can be made to move an object from one class (state) to another class. An action rule is composed of the changes, known as actions, that correlate with the change in the class value. The current work in action rule mining focuses on frequent, or highly occurring, action rules. The working assumption is that the end user is interested in class transitions that move a large number of objects from the initial class to the new (final) class with a high degree of confidence. Currently, very little work focuses on rare action rules; that is, classes which occur infrequently. In this paper, we provide a definition for a rare action rule and then propose a consequent-constraint based algorithm for generating these rare action rules.

Conducting digital forensic investigations in a big data distributed file system environment presents significant challenges to an investigator given the high volume of physical data storage space. Presented is an approach from which the Hadoop Distributed File System logical file space is mapped to the physical data location. This approach uses metadata collection and analysis to reconstruct events in a finite time series.

Distributed file systems present distinctive forensic challenges in comparison to traditional locally mounted file system volume. Storage device media can number in the thousands, and forensic investigations in this setting necessitate a tailored approach to data collection. The Hadoop Distributed File System (HFDS) produces and maintains partially persistent metadata that is pursuant with a logical volume, a file system, and file addresses on the centralized server. Hence, this research investigates the viability of using a residual central server digital artifact to generate a history model of the distributed file system. The history model affords an investigator a high-level perspective of low-level events to narrow investigative process obligations. The model is generated through set-theoretic relations of the file system essential data structure. Graph-theoretic ordering is applied to the events to provide a history model. The research contribution is a rapid reconstruction of the HDFS storage state transitions generating timelines for system events to forensically assess HDFS properties with conceptual similarity to traditional low-level file system forensic tool output. The results of this research provide a prototype tool, DFS 3 , for rapid and noninvasive data storage state timeline reconstruction in a big data distributed file system.

Association mining is the process of discovering relationships between items in a data set, where a group of items forms an itemset. A problem with the typical association mining approach is a large number of the generated frequent itemsets typically do not contain any items of interest to the user. Targeted association mining solves this by only deriving itemsets that include the items specified within a user's query. A popular targeted approach is the Itemset Tree, which consists of an index tree structure and algorithms for search and rule generation. Numerous enhancements to improve the Itemset Tree efficiency or extend its capability have been proposed. However, two major problems exist. First, the itemset generation process utilized by the Itemset Tree does not leverage the Apriori Principle, resulting in the unnecessary and costly generation of infrequent itemsets. Second, the Itemset Tree rule generation process has restrictions that prevent some rules containing the user query from being generated; as a result, the user misses useful information. In this paper, we offer a redesigned generation process resulting in faster itemset generation (milliseconds versus minutes) and expanded rulesets.

Action rule mining develops rules that describe which attributes should be changed in order to move an object from an undesired state to a desired state, with the understanding that some attributes cannot be changed. While such rules can be very useful for end-users, a limitation in prior work is the underlying assumption that the attributes of a dataset are discrete in nature. To address this limitation, we propose a method for generating action rules for objects described by continuously valued data. As part of the process, we developed a model for determining the effectiveness of the change, which permits more tailored recommendations for how to modify objects. Experimental results indicate that we can successfully create action rules for continuous valued data, and the use of automated tuning reduces the number of changes that must be performed to move an object from an undesired state to a desired state.

Reports and press releases highlight that security incidents continue to plague organizations. While researchers and practitioners' alike endeavor to identify and implement realistic security solutions to prevent incidents from occurring, the ability to initially identify a security incident is paramount when researching a security incident lifecycle. Hence, this research investigates the ability of employees in a Global Fortune 500 financial organization, through internal electronic surveys, to recognize and report security incidents to pursue a more holistic security posture. The research contribution is an initial insight into security incident perceptions by employees in the financial sector as well as serving as an initial guide for future security incident recognition and reporting initiatives.

In order for the Open Access (OA) to learning concept to a have wider impact in formal education, it is important that faculty members intent to adopt new educational innovations. However, little is known about which variables influence the intention of faculty members. Therefore, the purposes of this study are to empirically determine: 1) which of the characteristics of the educational innovation significantly influence the intention to adopt educational innovations, 2) which variables influence the readiness of faculty members intention to adopt educational innovations, and 3) how the characteristics of the innovations moderate the relationship between faculty readiness and intention to adopt the innovations. Participants of this study include 335 faculty members in ABET certified computer science and electrical engineering programs in the United States. The results show that ease of use is positively related to the intention of faculty members to adopt an educational innovation. We conclude that Open-Course Ware developers need to ensure that ease of use is emphasized in the Course Ware and they need to propagate these initially in institutions where faculty members have positive attitude to the Course Ware and care about student learning. In addition, a new method of identifying, building, and funding "open access grant" universities that develop easy-to-use educational innovations, make them available on an open access platform, and spread them widely by embedding agents in community colleges, schools, and other educational institutions is essential. Such an initiative may lead to wider adoption of MOOCS and other open access materials.

Substantial funds are invested in developing educational technologies with the goal that faculty members adopt and routinely use these technologies. This research establishes that there is a gap between the development and widespread adoption/use of these technologies. This paper investigates (1) which critical success factors (CSF) influence faculty members to adopt and routinely use technologies, and (2) whether the CSFs moderate the adoption process. Based on surveys of 335 computer science and electrical engineering faculty members, the research findings pinpoint the factors that influence the adoption and routine use of educational technologies. These results can be used by developers of educational technologies to create and establish plans to increase faculty awareness of and create positive attitudes towards the technologies.

Although substantial funding has been expended to develop new educational innovations, especially in engineering, few have found widespread acceptance in the classroom. Little is known regarding the interactions among the variables that influence successful dissemination and adoption. This research proposes a framework to show the interactions between innovation characteristics and the readiness of an educational organizations' faculty members, administration, and students to become aware, intend to adopt, adopt, and use educational innovations. One hundred eighty seven engineering education papers published during 2007-2012 were analyzed to identify the characteristics of educational innovations that influence their eventual dissemination. The researchers analyzed a sample of these articles (37) and synthesized the results to develop a framework. Interrater reliability among the researchers was 0.86. The proposed framework describes the interaction between characteristics of an educational innovation and an organizations' readiness to disseminate and adopt an innovation. The framework is then illustrated with examples. This framework provides a mechanism to link the outcomes, interventions, and context where the outcomes are the successful dissemination and adoption of educational innovations, the interventions are the characteristics of the innovation, and the context is the organizational readiness (readiness of faculty members, administrators, and students to adopt the innovation). For widespread dissemination and adoption to occur, educational innovators must incorporate positive characteristics in the innovations and implement them in organizations that are ready for such change.