Search results

Sarcasm is a linguistic expression that usually carries the opposite meaning of what is being said by words, thus making it difficult for machines to discover the actual meaning. It is mainly distinguished by the inflection with which it is spoken, with an undercurrent of irony, and is largely dependent on context, which makes it a difficult task for computational analysis. Moreover, sarcasm expresses negative sentiments using positive words, allowing it to easily confuse sentiment analysis models. This paper aims to demonstrate the task of sarcasm detection using the approach of machine and deep learning.

For the purpose of sarcasm detection, machine and deep learning models were used on a data set consisting of 1.3 million social media comments, including both sarcastic and non-sarcastic comments. The data set was pre-processed using natural language processing methods, and additional features were extracted and analysed. Several machine learning models, including logistic regression, ridge regression, linear support vector and support vector machines, along with two deep learning models based on bidirectional long short-term memory and one bidirectional encoder representations from transformers (BERT)-based model, were implemented, evaluated and compared.

The performance of machine and deep learning models was compared in the task of sarcasm detection, and possible ways of improvement were discussed. Deep learning models showed more promise, performance-wise, for this type of task. Specifically, a state-of-the-art model in natural language processing, namely, BERT-based model, outperformed other machine and deep learning models.

This study compared the performance of the various machine and deep learning models in the task of sarcasm detection using the data set of 1.3 million comments from social media.

This study aims to explore and empirically test variables influencing material delivery schedule inaccuracies?

A mixed-method case approach is applied. Explanatory variables are identified from the literature and explored in a qualitative analysis at an automotive original equipment manufacturer. Using logistic regression and random forest classification models, quantitative data (historical schedule transactions and internal data) enables the testing of the predictive difference of variables under various planning horizons and inaccuracy levels.

The effects on delivery schedule inaccuracies are contingent on a decoupling point, and a variable may have a combined amplifying (complexity generating) and stabilizing (complexity absorbing) moderating effect. Product complexity variables are significant regardless of the time horizon, and the item’s order life cycle is a significant variable with predictive differences that vary. Decoupling management is identified as a mechanism for generating complexity absorption capabilities contributing to delivery schedule accuracy.

The findings provide guidelines for exploring and finding patterns in specific variables to improve material delivery schedule inaccuracies and input into predictive forecasting models.

The findings contribute to explaining material delivery schedule variations, identifying potential root causes and moderators, empirically testing and validating effects and conceptualizing features that cause and moderate inaccuracies in relation to decoupling management and complexity theory literature?

The main aim of this study is to explore the appropriate factors in measuring COVID-19-induced supply chain disruptions and the impact of these disruptions on the economic vulnerability of small-scale farmers in Sri Lanka.

The findings revealed that most of the farmers have continued to cultivate even during the pandemic despite several challenges which affected their economic status. Therefore, it is concluded that COVID-19-induced transportation and demand disruptions exacerbated the economic vulnerability of small-scale farmers over the disruptions in supply and production.

The findings of this study are crucial for formulating novel policies to improve the sustainability of the Sri Lankan agricultural sector and alleviate the poverty level of Agri-communities in the countryside. As farming is a vital sector in the economy, increased attention ought to be given on facilitating farmers with government-encouraged loans or allowances for their financial stability. Further, the respective government authorities should develop programs for importing and distributing adequate quantities of fertilizers among all the farmers at controlled prices so that they can continue their operations without any interruption. Moreover, the government could engage in collaboratively work with private organizations to streamline the Agri-input supply process. There should be a government initiative for critical consideration of the issues of farming families and their continued motivation to engage in agriculture. Thus, farmers' livelihoods and agricultural prosperity could be upgraded through alternative Agri-inputs and marketing strategies, providing financial assistance, encouraging innovative technology, etc.

Despite the significance and vulnerability of the vegetable and fruit sector in Sri Lanka, there is a limitation in the empirical studies conducted on the supply chain disruptions caused by COVID-19 measures and their implications on the farmers' livelihood. Furthermore, previous empirical research has not employed adequate quantitative tools to analyze the situation or appropriate variables in evaluating COVID-19-induced disruptions. Hence, the current study explored the appropriate factors for measuring COVID-19-induced supply chain disruption using exploratory factor analysis. Then, the impact of those factors on the economic vulnerability of the small scale farmers was revealed through the ordinal logistics regression analysis.

This study aims to examine the relationship between medication for opioid use disorder (MOUD) among pregnant individuals, referral source, mental health, political affiliation and substance use policies in all 50 states in the USA.

This study describes MOUD receipt among pregnant people with an opioid use disorder (OUD) in 2018. The authors explored sociodemographic differences in MOUD receipt, referrals and co-occurring mental health disorders. The authors included a comparison of MOUD receipt among states that have varying substance use policies and examined the impact of these policies and the political affiliation on MOUD. The authors used multilevel binary logistic regression to examine effects of individual and state-level characteristics on MOUD.

Among 8,790 pregnant admissions with OUD, the majority who received MOUD occurred in the Northeast region (71.52%), and 14.99% were referred by the criminal justice system (n = 1,318). Of those who were self-referred, 66.39% received MOUD, while only 30.8% of referrals from the criminal justice system received MOUD. Those referred from the criminal justice system or who had a co-occurring mental health disorder were least likely to receive MOUD. The multilevel model showed that while policies were not a significant predictor, a state’s political affiliation was a significant predictor of MOUD.

The study has some methodological limitations; a state-level analysis, even when considering the individual factors, may not provide sufficient description of community-level or other social factors that may influence MOUD receipt. This study adds to the growing literature on the ineffectiveness of prenatal substance use policies designed specifically to increase the use of MOUD. If such policies are consistently assessed as not contributing to substantial increase in MOUD among pregnant women over time, it is imperative to investigate potential mechanisms in these policies that may not facilitate MOUD access the way they are intended to.

Findings from this study aid in understanding the impact that a political affiliation may have on treatment access; states that leaned more Democratic were more likely to have higher rates of MOUD, and this finding can lead to research that focuses on how and why this contributes to greater treatment utilization. This study provides estimates of underutilization at a state level and the mechanisms that act as barriers, which is a stronger assessment of how state-specific policies and practices are performing in addressing prenatal substance use and a necessary step in implementing changes that can improve the links between pregnant women and MOUD.

To the best of the authors’ knowledge, this is the first study to explore individual-level factors that include mental health and referral sources to treatment that lead to MOUD use in the context of state-level policy and political environments. Most studies estimate national-level rates of treatment use only, which can be useful, but what is necessary is to understand what mechanisms are at work that vary by state. This study also found that while substance use policies were designed to increase MOUD for pregnant women, this was not as prominent a predictor as other factors, like mental health, being referred from the criminal justice system, and living in a state with more Democratic-leaning affiliations.

The purpose of this paper is to identify the differences between Hungarian family-owned businesses (FOBs) and non-family-owned businesses (NFOBs) concerning the elements of SME competitiveness and financial performance.

The research covers the Hungarian data set of the Global Competitiveness Project (GCP, www.sme-gcp.org) of 738 (data collection between 2018 and 2020) non-listed SMEs, of which 328 were FOBs. The study uses the comprehensive, multidimensional competitiveness measurement of the GCP built on the resource-based view (RBV) and the configuration theory. Financial performance was captured with two composite indicators: short-term and long-term financial performance (LTFP). The comparative analysis between FOBs and NFOBs was conducted using binary logistic regression.

The results show that FOBs are more prone to focusing on local niche markets with higher longevity and LTFP than NFOBs. However, FOBs have lower innovation intensity and less organised administrative procedures. The most contradicting finding is that the FOBs’ higher LTFP is accompanied by significantly lower competitiveness than in the case of NFOBs.

This study goes beyond other GCP studies by including composite financial performance measures among the variables examined. The combination of performance-causing (resources and capabilities) and performance-representing (financial performance) variables provides a better understanding of the non-listed SMEs in terms of family ownership. The results help academia to enrich the RBV-competitiveness, the non-listed SME management and finance literature, and policymakers to design business development and support schemes. They also show future entrepreneurs the impact of family ownership on entrepreneurial success.

Drawing on COR theory and based on a person-centered approach, this study aims to explore profiles of both leadership behavior (transformational leadership, abusive supervision) and well-being indicators (cognitive irritation, emotional exhaustion). Additionally, we consider whether certain resource-draining (work intensification) and resource-creating factors (leader autonomy, psychological contract fulfillment) from the leaders' work context are related to profile membership.

The profiles are built using LPA on data from 153 leaders and their 1,077 followers. The relationship between profile membership and correlates from the leaders' work context is examined using multinomial logistic regression analyses.

LPA results in an interpretable four-profile solution with the profiles named (1) Good health – constructive leading, (2) Average health – inconsistent leading, (3) Impaired health – constructive leading and (4) Impaired health – destructive leading. The two groups with the highest sample share – Profiles 1 and 3 – both show highly constructive leadership behavior but differ significantly in their well-being indicators. The regression analyses show that work intensification and psychological contract fulfillment are significantly related to profile membership.

The person-centered approach provides a more nuanced view of the leadership behavior – leader well-being relationship, which can address inconsistencies in previous research. In terms of practical relevance, the person-centered approach allows for the identification of risk groups among leaders for whom organizations can provide additional resources and health-promoting interventions.

The purpose of this study is to estimate the effects of gender composition in the workplace on employee job satisfaction and commitment.

The data were collected on both the organizational and employee levels at three different points in time in organizations with more than 70 employees. Multi-level mixed-effects ordered logistics regressions were used to account for the multi-level nature of the data and the ordered nature of the dependent variables.

Employees in gender-balanced workplaces show higher levels of job satisfaction and commitment than those in female-dominated or male-dominated workplaces. The relationship is also based on the gender of the individual, as men show a significantly lower level of both job satisfaction and commitment when working in male-dominated workplaces than others, while for women, the effect is only significant for commitment.

Aiming for a balance in the gender composition of the workplace may improve employee attitudes, especially for men. The results also indicate that further research is warranted into why job satisfaction and commitment are significantly lower among men in male-dominated workplaces.

The relationship between gender and job satisfaction and commitment is well established, but less is known about the effects of gender composition on job satisfaction and commitment. Previous papers have focused on job satisfaction. This paper extends prior studies by estimating the effects of gender composition on both job satisfaction and commitment using multi-level regressions on a rich dataset.

This study aims to examine the potentially important effects of academic embeddedness on college of business student retention and performance as well as the mediating effects of self-efficacy on the academic embeddedness student outcomes relationships. Improvements in student retention and performance reduce costs for students and universities and lead to higher incomes for graduates.

Data were gathered from students in an entry-level business course at a public university in a rural western state. Approximately 45% of the students were female, and the average age of participants was 20 years old. A survey was administered midsemester to gather data on academic embeddedness and self-efficacy. Retention was indicated by a student enrolling in a business course in a subsequent semester. Performance was measured using end-of-semester course grades. Logistic and linear regression as well as mediation analysis were used to test the hypotheses.

Academic embeddedness was found to positively predict both retention and performance, while self-efficacy was found to positively mediate the academic embeddedness retention relationship. The direct effect of embeddedness on performance was not found when controlling for self-efficacy.

Student retention and performance are important to both students and academic administrators. The findings of this study suggest that retention and performance can both be improved by focusing on factors that more strongly embed students to their colleges.

Embeddedness has been found to have high predictive validity in the employment context. This is one of the first studies to consider the effects of embeddedness in the academic context.

In the literature, evidence is to be found of the positive effect of high-performance work systems (HPWSs) on innovation in firms. However, innovation is enabled by not only human resources but also digital technology, and scholars have called for further investigation into the interplay between digital technology and HRM systems. Drawing on signalling theory and HPWSs research, the purpose of this study is to explore the moderating role of digital technologies in the relationship between HPWSs and innovation in the firm and consider employee participation as an additional conditioning factor.

This study uses data from the European Company Suvery 2019 administered in a sample of more than 20,000 European establishments and applies logistic regression with a three-way interaction.

HPWSs underpin product and process innovation. Moreover, this study shows that in firms with low levels of employee participation, digital technology enhances the effect of HPWSs on innovation, while in firms with high levels of employee participation, this effect is reduced.

This study enriches the scholarly discussion about the link between HPWSs and innovation in the firm, by investigating in theoretical and empirical terms the moderating effect of digital technology, underlining that either positive or negative synergistic effects are possible. By adding employee participation to the analysis, the authors cast light on an important boundary condition for understanding when the synergic effects become more prominent. This intends to respond to recent calls from scholars and practitioners for more insight into the precise nature of the synergies between HPWSs and digital technology on innovation in the firm, with important implications for management.

Coronavirus disease 2019 (COVID-19) pandemic has led to unprecedented mental health repercussions in the lives of every individual including university students. Therefore, study on students’ psychological state and its associated factors during the pandemic are of importance. This study aims to discuss the aforementioned issue.

An online survey was done on a total of 207 university students of Pakistan to collect information on socio-demographic characteristics, concerns or fears amidst COVID-19 and mental distress. Validated tools; Perceived Stress Scale (PSS), Generalized Anxiety Disorder Scale (GAD-7) and Patient Health Questionnaire (PHQ-9)-Depression were used to assess stress, anxiety and depression, respectively.

Around 14% of the university students were experiencing severe stress and anxiety, while 8.2% had severe depression. The authors found that stress among university students was related to psychiatric illness or symptoms (OR = 5.1: 1.1, 22.9) and unpredictability due to the pandemic (OR = 3.7: 1.2, 11.2). The significant determinants of anxiety were psychiatric illness/symptoms (OR = 6.6: 3.4, 12.9), implementation of public health measures (OR = 3.7: 1.1, 11.6), employed mothers (OR = 2.4: 1.1, 5.0) and lack of support from university administration (OR = 2.2: 1.0, 5.0). While the factors associated with depression included psychiatric illness or symptoms (OR = 8.4: 3.3, 21.5), unpredictability due to pandemic (OR = 6.8: 2.2, 20.7), impaired social support system (OR = 3.7: 1.3, 10.4) and studying without a scholarship (OR = 2.1: 1.0, 4.4).

These findings call for an urgent need to develop appropriate interventions and educational programs that could address the psychological needs of students.

The study directs the role of university and faculty in dealing the mental health needs of the student in COVID-19 pandemic time.

Educational programs are important that could address the psychological needs of students in COVID-19 pandemic.

University students reported mental distress during COVID-19 pandemic which shows that younger people are at risk of COVID-19 repercussions. Moreover, several stressors (i.e. impaired social support system and lack of support from universities) were revealed that could be mitigated by implementing appropriate strategies.