Search results

The purpose of this paper is to alleviate the problem of poor robustness and over-fitting caused by large-scale data in collaborative filtering recommendation algorithms.

Interpreting user behavior from the probabilistic perspective of hidden variables is helpful to improve robustness and over-fitting problems. Constructing a recommendation network by variational inference can effectively solve the complex distribution calculation in the probabilistic recommendation model. Based on the aforementioned analysis, this paper uses variational auto-encoder to construct a generating network, which can restore user-rating data to solve the problem of poor robustness and over-fitting caused by large-scale data. Meanwhile, for the existing KL-vanishing problem in the variational inference deep learning model, this paper optimizes the model by the KL annealing and Free Bits methods.

The effect of the basic model is considerably improved after using the KL annealing or Free Bits method to solve KL vanishing. The proposed models evidently perform worse than competitors on small data sets, such as MovieLens 1 M. By contrast, they have better effects on large data sets such as MovieLens 10 M and MovieLens 20 M.

This paper presents the usage of the variational inference model for collaborative filtering recommendation and introduces the KL annealing and Free Bits methods to improve the basic model effect. Because the variational inference training denotes the probability distribution of the hidden vector, the problem of poor robustness and overfitting is alleviated. When the amount of data is relatively large in the actual application scenario, the probability distribution of the fitted actual data can better represent the user and the item. Therefore, using variational inference for collaborative filtering recommendation is of practical value.

It is of great significance to study the influence of subgrade filling on permafrost temperature field in permafrost area for the smooth construction and safe operation of railway.

The paper builds up the model for the hydrothermal coupling calculation of permafrost using finite element software COMSOL to study how permafrost temperature field changes in the short term after subgrade filling, on which basis it proposes the method of calculation for the concave distortion of freezing front in the subgrade-covered area.

The results show that the freezing front below the subgrade center sinks due to the thermal effect of subgrade filling, which will trigger hydrothermal erosion in case of sufficient moisture inflows, leading to the thawing settlement or the cracking of the subgrade, etc. The heat output of soil will be hindered the most in case of July filling, in which case the sinking and the distortion of the freezing front is found to be the most severe, which the recovery of the permafrost temperature field, the slowest, constituting the most unfavorable working condition. The concave distortion of the freezing front in the subgrade area increases with the increase in temperature difference between the filler and ground surface, the subgrade height, the subgrade width and the volumetric thermal capacity of filler, while decreases with the increase of the thermal conductivity of filler. Therefore, the filler chose for engineering project shall be of small volumetric thermal capacity, low initial temperature and high thermal conductivity whenever possible.

The concave distortion of the freezing front under different working conditions at different times after filling can be calculated using the method proposed.

Despite retail digitisation and research efforts focussed on online and omnichannel shopping, there is insufficient knowledge regarding retail patronage formation in the grocery category, where in-store sales dominate. This study analyses the retail patronage formation in grocery in-store fill-in shopping.

The authors designed a questionnaire to measure retail patronage behaviour, consumer satisfaction (CS), store attributes evaluation and e-retail usage. Then, the authors analysed the path structure for retail patronage behaviour formation using structural equation modelling. Additionally, they performed a mediation analysis using the bootstrap method and a moderation analysis based on a chi-square difference test.

This study provides three main findings. First, the authors' model has two ways to increase Share-of-Wallet (SOW). One is to increase Share-of-Visits (SOV) and another is to increase CS amongst non-users of e-retailing. Second, the results of the moderation analysis suggest the influence of customers' use or non-use of e-retailing on SOW formation. Third, service evaluation plays an interesting role in the overall model: the lower the assessment of service, the higher the SOV; the higher the evaluation of service, the greater the CS; the greater the CS, the higher the SOV.

The authors proposed the framework for the relative retail patronage formation in grocery fill-in shopping to examine the relationship between two relative patronage indicators (SOW and SOV) in the path structure and the mediating effect of CS and the moderating effect of e-retailing usage on retail patronage formation.

This study aims to investigate the merits of the unified model of task-specific motivation (UMTM) in predicting transfer of training and to investigate (relationships between) changes in UMTM components over time. In doing so, this study takes the multidimensionality of transfer motivation into account.

The authors collected data among 514 employees of the judiciary who filled in the UMTM questionnaire directly after the training and after three weeks. The data were analyzed by means of structural equation modelling.

The outcomes show that transfer motivation predicts transfer intention and transfer of training over time. Moreover, the study shows that (change in) transfer motivation is predicted by (change in) personal and contextual factors identified by the UMTM as antecedents of motivation.

This study describes the first longitudinal evaluation of the UMTM in the literature and shows its applicability for predicting transfer of training. It is also one of the few studies that investigate transfer motivation multidimensionally and the role it plays for transfer of training. As such, this study informs other transfer of training models about the nature of transfer motivation and how transfer of training could be predicted.

An Autodesk Ember three-dimensional (3D) printer was used to print optical components from Clear PR48 photocurable resin. The cured PR48 was characterized by the per cent of light transmitted and the index of refraction, which was measured with a prism spectrometer. Lenses and diffraction gratings were also printed and characterized. The focal length of the printed lenses agreed with predictions based on the thin lens equation. The periodicity and effective slit width of the printed gratings were determined from both optical micrographs and fits to the Fraunhofer diffraction equation. This study aims to demonstrate the advantages offered by a layer-by-layer DLP printing process for the manufacture of optical components for use in the visible region of the electromagnetic spectrum.

A 3D printer was used to print both lenses and diffraction gratings from Standard Clear PR48 photocurable resin. The manufacturing process of the lenses and the diffraction gratings differ mainly in the printing angle with respect to the printer x-y-axes. The transmission diffraction gratings studied here were manufactured with nominal periodicities of 10, 25 and 50 µm. The aim of this study was to optically determine the effective values for the distance between slits, d, and the effective width of the slits, w, and to compare these values with the printed layer thickness.

The normalized diffraction patterns measured in this experiment for the printed gratings with layer thickness of 10, 25 and 50 µm are shown by the solid dots in Figures 8(a)-(c). Also shown as a red solid line are the fits to the experimental diffraction data. The effective values of d and w obtained from fitting the data are compared to the nominal layer thickness of the printed gratings. The effective distance between slits required to fit the diffraction patterns are well approximated by the printed layer thickness to within 14, 4 and 16 per cent for gratings with a nominal 10, 25 and 50 µm layer thickness, respectively.

Chromatic aberration is present in all polymer lenses, and the authors have not attempted to characterize it in this study. These materials could be used for achromatic lenses if paired with a crown-type material in an achromatic doublet configuration, because this would correct the chromatic aberration issues. It is worthwhile to compare the per cent transmission in cured PR48 resin (approximately 80 per cent) to the percent transmission found in common optical materials like BK7 (approximately 92 per cent) over the visible region. The authors attribute the lower transmission in PR48 to a combination of surface scattering and increased absorption. At the present time, the authors do not know what fraction of the lower transmission is related to the surface quality resulting from sample polishing.

There are inherent limitations to the 3D manufacturing process that affect the performance of lenses. Approximations to a curved surface in the design software, the printing resolution of the Autodesk Ember printer and the anisotropy due to printing in layers are believed to be the main issues. The performance of the lenses is also affected by internal imperfections in the printed material, in particular the presence of bubbles and the inclusion of debris like dust or fibers suspended in air. In addition, the absorption of wavelengths in the blue/ultraviolet produces an undesirable yellowing in any printed part.

One of the most interesting results from this study was the manufacture of diffraction gratings using 3D printing. An analysis of the diffraction pattern produced by these printed gratings yielded estimates for the slit periodicity and effective slit width. These gratings are unique because the effective slit width fills the entire volume of the printed part. This aspect makes it possible to integrate two or more optical devices in a single printed part. For example, a lens combined with a diffraction grating now becomes possible.

Drawing on the signaling theory and technology acceptance model, the main purpose of this study was to predict prospective employees' intentions to apply for jobs in a firm, with a special focus on the mediating role of attitudes toward corporate websites and the moderating role of perceived value fit.

Collecting data from a convenient sample of 318 prospective job candidates, the research hypotheses were tested using structural equation modeling (SEM) with AMOS (version 24) and SPSS Process Macro (version 3.4).

The test results revealed that prospective employees' attitudes toward corporate websites partially mediate the association of corporate reputations, perceived ease of use and perceived usefulness with their intentions to apply for jobs in an organization. Noticeably, perceived value fit moderated the perceived usefulness–application intentions link in such a way that the impact of perceived usefulness on intentions to apply appears higher for individuals with a low level (than a high level) of perceived value fit.

Consistent with the research findings, a notable theoretical contribution and practical implications for HR professionals have been discussed. This paper ends with outlining some limitations and future research directions.

Despite having the salient buffering effects of perceived value fit on the applicant attraction process, empirical study on this theoretical phenomenon is still sparse in a pre-employment context. This may be the first study that demonstrates under what circumstances prospective employees' job pursuit intentions could be optimized in respect of their perceived value fit within a single framework comprised of two theories.

I aimed to obtain a deeper insight into the link between supplier involvement in product development (SIPD), supplier relationship resilience and company performance.

To collect data, a survey among 500 Polish manufacturing companies was conducted. I used quantitative methods (structural equation modeling) to test several research hypotheses referring to a single supplier–customer relationship. Thanks to the use of multi-construct measurement of SIPD and supplier relationship resilience, the study provides detailed research results on the topic.

Collaborative practices implemented during SIPD increase procurement flexibility and decrease redundancy in the relationship with the involved supplier. Communication during SIPD increases supplier flexibility and procurement flexibility. Increased supplier flexibility and increased procurement flexibility in the relationship with the involved supplier as well as collaborative practices during SIPD positively impact company performance. I confirmed the indirect effect between communication during SIPD and company performance when the mediators are supplier flexibility and procurement flexibility. Decreased redundancy in relationship with involved supplier does not impact company performance.

Supply chain managers need to rethink SIPD practice to effectively ensure supply chain resilience (SCRES), especially in the face of the contemporary global crisis and black swans affecting the supplier base. My article provides important managerial insights into drivers of SCRES and company performance.

To the best of my knowledge, this research is among the first to conclude that SIPD does not have an unequivocally positive or direct impact on supplier relationship resilience. The research fills the gap by analyzing the impact of SIPD on two main SCRES elements. The study examines supplier relationship resilience, understood as flexibility and redundancy elements, in a single supplier–buyer relationship perspective. Thus, the presented considerations go beyond the traditional understanding of flexibility and redundancy in supplier relationship management, that is through the prism of double or multi sourcing and having back up-suppliers.

This study aims to analyze the effect of service quality on trust, satisfaction and loyalty by adopting two models, namely, conventional service quality model from Parasuraman and information systems (IS) success model from Delone and McLean.

Respondents of this study were users of shared-motorcycle services who filled out a complete questionnaire totaling 507. This research used a second-order structural equation model. All question items had quite high reliability and validity based on the result of confirmatory factor analysis with a value of average variance extracted and composite reliability which was higher than 0.70. The goodness of fit was quite good with the values x²/df = 2.493, incremental fit index = 0.921, Tucker-Lewis index = 0.917, comparative fit index = 0.921 and root-mean-square error of approximation = 0.054.

Online and offline ride-sharing services reveal a strong and positive influence on trust and satisfaction. Trust reveals a strong and positive influence on satisfaction and loyalty. Finally, satisfaction reveals a strong and positive influence on loyalty. The research in general shows that the quality of offline service is more influential than the quality of online service in the case of ride-sharing service provided by two companies in Indonesia.

The sampling frame of the research was diverse, including students of various collages and junior high schools, various private company workers and government employees. So, the results cannot be generalized to all populations especially to all Indonesian customers. It is recommended to increase the number of samples by focusing on the community groups of customers of public motorbikes, so that these groups can be compared. Next, the research finds that both service quality based on IS and service quality models reveal a strong and positive influence on loyalty both directly and indirectly.

The research uses respondents who use motorcycle services both online and offline. The findings of the research are important for online and offline ride-sharing motorbike service providers. They have to maintain their excellent services to the customers.

In a dynamic environment, understanding the consumer’s behavior in an online market is critical for the development of online retailers’ strategies. In Egypt, although the number of internet users is growing rapidly, online shopping is at an early stage of development. In addition, there is a lack of knowledge about the behavior of Egyptians toward online shopping and the factors that influence their behavior.

To demonstrate the applicability of the technology acceptance model (TAM), this research applies an extended TAM to analyze the factors that impact the Egyptian consumers’ online shopping behavior. Structural equation modeling is used for fitting the data and testing the hypothesis.

Findings revealed that perceived enjoyment, perceived ease of use, social norm and perceived risk have significant influences on the respondents to shop online; while the site language effect was insignificant.

This study has mainly four limitations. The first concerning the sample, it should cover diversified areas other than Cairo and should include respondents from different education levels. Future studies may consider using a national sample of current and potential online shoppers to be able to generalize the results. Second, the questionnaire should include questions about the levels of education and the level of income to analyze if they influence actual online shopping. Third, although the benefits of using online surveys such as reducing the costs and time and collecting data more accurately compared to paper-pencil surveys, the low response rate may lead to non-response bias. Finally, constructs such as website content, design, and response could be added to the TAM to examine different aspects of website functionalities on the actual shopping.

Considering that online shopping is still at the early stage of development in Egypt, there is a need to examine the factors that influence the behavior of Egyptian consumers while adopting online shopping. To fill in this gap, this study is analyzing the factors impacting the consumers’ decisions to shop online by examining a sample of students from the Faculty of Economics and Political Science, Cairo University. This study will contribute to providing an empirical application of the TAM on the Egyptian consumers in addition to, enhancing the knowledge of electronic retailers about the factors that drive consumers toward shopping online.

The purpose of this paper was to establish the psychometric properties of the Asian Family Characteristics Scale (AFCS) in the Thai population.

The 30-item AFCS originally developed in the Malay language was translated into Thai. Thai (n = 384) and Malay (n = 500) participants in Study 1 responded to the AFCS in their respective languages. The data were subjected to a confirmatory factor analysis with a measurement invariance test. In Study 2, Thai participants (n = 495) filled out the AFCS and Chulalongkorn Family Index, International Personality Item Pool-NEO, Self-Compassion Scale, Depression, Anxiety, and Stress Scale (DASS-21) and Satisfaction with Life Scale.

Study 1 showed that the measurement model of the Thai AFCS fit the data from the Thai population. The measurement invariance test confirmed that the structure and meaning of the AFCS are equivalent across the Thai and Malay samples. Study 2 demonstrated the AFCS's convergent validity by showing that the AFCS score had a positive correlation with the Chulalongkorn Family Inventory, self-compassion, agreeableness, conscientiousness, satisfaction with life and a negative correlation with neuroticism, depression, anxiety and stress. The AFCS's discriminant validity was supported by nonsignificant correlations with extraversion and openness to experience.

This paper is an attempt to develop a family characteristic measure specifically for the Asian population. The results provide empirical evidence for measurement invariance and validity of the scale in another Asian language, enhancing its cross-cultural generalizability.