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.

The purpose of this paper is to combine the entropy weight method with the cloud model and establish a fire risk assessment method for airborne lithium battery.

In this paper, the fire risk assessment index system is established by fully considering the influence of the operation process of airborne lithium battery. Then, the cloud model based on entropy weight improvement is used to analyze the indexes in the system, and the cloud image is output to discuss the risk status of airborne lithium batteries. Finally, the weight, expectation, entropy and hyperentropy are analyzed to provide risk prevention measures.

In the risk system, bad contact of charging port, mechanical extrusion and mechanical shock have the greatest impact on the fire risk of airborne lithium battery. The fire risk of natural factors is at a low level, but its instability is 25% higher than that of human risk cases and 150% higher than that of battery risk cases.

The method of this paper can evaluate any type of airborne lithium battery and provide theoretical support for airborne lithium battery safety management.

After the fire risk assessment is completed, the risk cases are ranked by entropy weight. By summarizing the rule, the proposed measures for each prevention level are given.

Motivated by the problems that the positioning error of strap-down inertial navigation system (SINS) accumulates over time and few sensors are available for midwater navigation, this paper aims to propose a self-aided SINS scheme for the spiral-diving human-occupied vehicle (HOV) based on the characteristics of maneuvering pattern and SINS error propagation.

First, the navigation equations of SINS are simultaneously executed twice with the same inertial measurement unit (IMU) data as input to obtain two sets of SINS. Then, to deal with the horizontal velocity provided by one SINS, a delay-correction high-pass filter without phase shift and amplitude attenuation is designed. Finally, the horizontal velocity after processing is used to integrate with other SINS.

Simulation results indicate that the horizontal positioning error of the proposed scheme is less than 0.1 m when an HOV executes spiral diving to 7,000 meters under the sea and it is inherently able to estimate significant sensors biases.

The proposed scheme can provide a precise navigation solution without error growth for spiral-diving HOV on the condition that only IMU is required as a navigation sensor.

This paper aims to present a kinematics performance analysis and control for a continuum robot based on a dynamic model to achieve control of the robot.

To analyze the motion characteristics of the robot, its kinematics model is derived by the geometric analysis method, and the influence of the configuration parameters of the robot on workspace is investigated. Moreover, the dynamic model is established by the principle of virtual work to analyze the mapping relationship among the bending shape, the forces/torques applied to the robot. To achieve better control of the robot, a control strategy for continuum robot based on the dynamic model is put forward.

Results of the simulations and experiments verify the proposed continuum structure and motion model, the maximum position error is 5.36 mm when the robot performs planar bending motion and the average position error of the robot in spatial circular motion is 5.84 mm. The proposed model can accurately describe the deformation movement of the robot and realize its motion control with a few position errors.

The kinematics analysis and control model proposed in this paper can achieve precise control of the robot, which can be used as a reference for the motion planning and shape reconstruction of continuum robot.

In the paradox of personalized services and privacy risks, what factors influence users’ decisions is considered an interesting issue worth exploring. The current study aims to empirically explore privacy behavior of social media users by developing a theoretical model based on privacy calculus theory.

Privacy risks, conceptualized as natural risks and integrated risks, were proposed to affect the intention of privacy disclosure and protection. The model was validated through a hybrid approach of structural equation modeling (SEM)-artificial neural network (ANN) to analyze the data collected from 527 effective responses.

The results from the SEM analysis indicated that social interaction and perceived enjoyment were strong determinants of perceived benefits, which in turn played a dominant role in the intention to disclose the privacy in social media. Similarly, trust and privacy invasion experience were significantly related to perceived risks that had the most considerable effect on users’ privacy protection intention. And the following ANN models revealed consistent relationships and rankings with the SEM results.

This study broadened the application perspective of privacy calculus theory to identify both linear and non-linear effects of privacy risks and privacy benefits on users’ intention to disclose or protect their privacy by using a state-of-the-art methodological approach combining SEM and ANN.

The purpose of this paper is to explain how sociability and usability enhanced members’ satisfaction, and how such satisfaction in turn, influenced their continuance intention of knowledge sharing in academic virtual communities (AVCs).

Drawing on social capital theory and technology acceptance model, this study proposed a theoretical socio-technical model, and the partial least squares method is used to examine the proposed model, based on data collected from 431 subjects in a well-known academic community in China (i.e. ScienceNet).

Both sociability and usability were important to improve members’ satisfaction with knowledge sharing in AVCs. Specifically, social interaction ties, trust, reciprocity, shared vision, perceived ease of use and perceived usefulness are antecedents of members’ satisfaction, which in turn positively affects their continuance intention of knowledge sharing in AVCs.

This study provided insights that can help AVCs’ administrators develop effective strategies that could encourage continued knowledge sharing behavior through promoting members’ satisfaction.

While the socio-technical framework has mainly been used to study initial adoption and participation of knowledge sharing. This study proposed a socio-technical model to move a step forward by explaining the exact roles of sociability and usability in terms of promoting members’ satisfaction and identifying its critical effect on their continuance intention to share knowledge in AVCs, leading to a more comprehensive picture of members’ satisfaction and continuance intention of knowledge sharing in AVCs.

This study aims to evaluate the influence of oleic acid (OA)-capped Al₂O₃ nanoparticles on the tribological performance of conventional lube oil. The goal is to determine the optimal nanoparticle concentration that enhances lubricant efficiency by reducing friction and wear.

The research involved preparing nanolubricants with four different concentrations of Al₂O₃ nanoparticles: 0.05, 0.1, 0.25 and 0.5 wt.%. Tribological performance was assessed using a four-ball tribotester, which measured the coefficient of friction (COF) and wear scar diameter (WSD) under standardized testing conditions.

The experimental results revealed that the nanolubricant containing 0.1 wt.% OA-Al₂O₃ nanoparticles exhibited the most significant improvement in tribological performance. This formulation achieved a 38.84% reduction in COF and a 23.87% reduction in WSD compared to the base lubricant. These findings demonstrate the effectiveness of incorporating OA-capped Al₂O₃ nanoparticles in reducing friction and wear, thereby enhancing the overall performance of conventional lubricants.

This study demonstrates the benefits of OA-capped Al₂O₃ nanoparticles in lubricants, including a 38.84% reduction in COF and a 23.87% reduction in WSD. By systematically analyzing different nanoparticle concentrations, it identified that 0.1% by weight of nanoparticles is the most effective formulation for reducing friction and wear.

The peer review history for this article is available at: https://publons.com/publon/10.1108/ILT-06-2024-0236/

The purpose of this paper is to enhance the characteristics of TiO₂ nanoparticles (NPs). Because these NPs stick together easily and are difficult to distribute evenly, they cannot be used extensively in lubricating oils. Altering TiO₂ was recommended as an alternate way for making NPs simpler to disperse.

Through dielectric barrier discharge plasma (DBDP)-assisted ball mill diagnostics and modeling of molecular dynamics, TiO₂@PEG-400 NPs were produced using the DBDP-assisted ball mill. The NPs’ microstructure was examined using FESEM, TEM, XRD, FT-IR and TG-DSC. Using the CFT-1 reciprocating friction tester, the tribological properties of TiO₂@PEG-400 NPs as base oil additives were studied. EDS and XPS were used to examine the surface wear of the friction pair.

Tribological properties of the modified NPs are vastly superior to those of the original NPs, and the lipophilicity value of TiO₂ NPs was improved by 200%. It was determined through tribological testing that TiO₂@PEG-400’s exceptional performance might be attributable to a chemical reaction film made up of TiO₂, Fe₂O₃, iron oxide and other organic chemicals.

This work describes an approach for preventing the aggregation of TiO₂ NPs by coating their surface with PEG-400. In addition, the prepared NPs can enhance the tribological performance of lubricating oil. This low-cost, high-performance lubricant additive has tremendous promise for usage in marine engines to minimize operating costs while preserving navigational safety.

Concerns over the pollution caused by internal combustion vehicles have increased owing to population and industrialization increment. Addressing the confrontations, the demand for electric vehicles (EVs) as a combustion engine substitute became necessary in responding to environmental worries from internal combustion. The development of bio lubricant in lubricating the sliding parts of EVs is required to maintain the sustainability idea and to improve the system performance, which this research tends to explore.

In this research, the enhancement of base Jatropha oil was done using polytetrafluoroethylene (PTFE) and hexagonal boron nitrate (h-BN) as additives. Different characterization was conducted on the new formulation to ascertain its anticorrosion tendency. The wear and friction behavior of the formulations on the tribo-pairs surfaces in contact were investigated using ball on flat tribometer to determine their tribological responsiveness as mineral lubricant alternative. To explore the surface topography, surface profilometer, scanning electron microscope and energy dispersive X-ray investigations were PTFE, lubrication and EV carried out.

The test’s input parameters were EVs’ usual load and sliding speed, and the addition concentrations for PTFE were 0.3 Wt.%, 0.4 Wt.%, 0.5 Wt.% and 0.6 Wt.%, whereas h-BN were 0.4 Wt.%, 0.8 Wt.% and 1.2 Wt.%, respectively. The study on corrosion demonstrated resistance when applied PTFE and h-BN additives in Jatropha oil. The analysis revealed that 0.5 Wt.% PTFE + 0.8 Wt.% h-BN concentrations significantly improved the tribological characteristics when compared to the base Jatropha oil. The application of formulations yielded percentage reduction of 8.67%, 10.98%, 7.34% and 7.35%, respectively, for 0.5% poly + 0.5% h-BN, 0.5% poly + 0.6% h-BN, 0.5% poly + 0.7% h-BN, 0.5% poly + 0.8% h-BN against base Jatropha oil under 20 N.

The formulation of PTFE and h-BN for electric transmission with wear and friction effects was accomplished in this paper. The mechanism of particle diffusing at the sliding contact on tribological behavior could be examined based on the created model of operation.

The peer review history for this article is available at: https://publons.com/publon/10.1108/ILT-07-2023-0235/