Search results

This paper aims to propose an incremental graph indexing method based on probabilistic inferences in Bayesian network (BN) for approximate nearest neighbor search (ANNS) that adds unindexed queries into the graph index incrementally.

This paper first uses the attention mechanism based graph convolutional network to embed a social network into the low-dimensional vector space, which could improve the efficiency of graph index construction. To add the unindexed queries into the graph index incrementally, this study proposes to learn the rule-based BN from social interactions. Thus, the dependency relations of unindexed queries and their neighbors are represented, and the probabilistic inferences in BN are then performed.

Experimental results demonstrate that the proposed method improves the search precision by at least 5% and search efficiency by 10% compared to the state-of-the-art methods.

This paper proposes a novel method to construct the incremental graph index based on probabilistic inferences in BN, such that both indexed and unindexed queries in ANNS could be addressed efficiently.

Product returns are a major challenge for e-businesses as they involve huge logistical and operational costs. Therefore, it becomes crucial to predict returns in advance. This study aims to evaluate different genres of classifiers for product return chance prediction, and further optimizes the best performing model.

An e-commerce data set having categorical type attributes has been used for this study. Feature selection based on chi-square provides a selective features-set which is used as inputs for model building. Predictive models are attempted using individual classifiers, ensemble models and deep neural networks. For performance evaluation, 75:25 train/test split and 10-fold cross-validation strategies are used. To improve the predictability of the best performing classifier, hyperparameter tuning is performed using different optimization methods such as, random search, grid search, Bayesian approach and evolutionary models (genetic algorithm, differential evolution and particle swarm optimization).

A comparison of F1-scores revealed that the Bayesian approach outperformed all other optimization approaches in terms of accuracy. The predictability of the Bayesian-optimized model is further compared with that of other classifiers using experimental analysis. The Bayesian-optimized XGBoost model possessed superior performance, with accuracies of 77.80% and 70.35% for holdout and 10-fold cross-validation methods, respectively.

Given the anonymized data, the effects of individual attributes on outcomes could not be investigated in detail. The Bayesian-optimized predictive model may be used in decision support systems, enabling real-time prediction of returns and the implementation of preventive measures.

There are very few reported studies on predicting the chance of order return in e-businesses. To the best of the authors’ knowledge, this study is the first to compare different optimization methods and classifiers, demonstrating the superiority of the Bayesian-optimized XGBoost classification model for returns prediction.

The purpose of this paper is to develop a novel optimization method that can improve the convergence of the multi-material topology.

In the proposed method, the optimization procedure is divided into two steps. In the first step, a global search is performed to probabilistically determine the material distribution of multi-segmented magnets. In the second step, the design area is limited and a local search is performed to determine the detailed magnet shape.

Because the first optimization step determines the arrangement of the magnetization vectors according to the rotational position, as in a d-axis flux concentration orientation, the optimal solution can be obtained with a smaller volume of magnets than the conventional method.

Because a few case studies are considered in this paper, additional verification is required, such as application to different types of motors, to clarify scalability.

The solution obtained using the proposed method has a smaller amount of magnet than the solution obtained using the conventional method and can fully satisfy the average torque constraint.

The proposed method differs from the conventional method in that the material distribution is determined according to the probability function in the first optimization step.

Current reinforcement learning (RL) algorithms are facing issues such as low learning efficiency and poor generalization performance, which significantly limit their practical application in real robots. This paper aims to adopt a hybrid model-based and model-free policy search method with multi-timescale value function tuning, aiming to allow robots to learn complex motion planning skills in multi-goal and multi-constraint environments with a few interactions.

A goal-conditioned model-based and model-free search method with multi-timescale value function tuning is proposed in this paper. First, the authors construct a multi-goal, multi-constrained policy optimization approach that fuses model-based policy optimization with goal-conditioned, model-free learning. Soft constraints on states and controls are applied to ensure fast and stable policy iteration. Second, an uncertainty-aware multi-timescale value function learning method is proposed, which constructs a multi-timescale value function network and adaptively chooses the value function planning timescales according to the value prediction uncertainty. It implicitly reduces the value representation complexity and improves the generalization performance of the policy.

The algorithm enables physical robots to learn generalized skills in real-world environments through a handful of trials. The simulation and experimental results show that the algorithm outperforms other relevant model-based and model-free RL algorithms.

This paper combines goal-conditioned RL and the model predictive path integral method into a unified model-based policy search framework, which improves the learning efficiency and policy optimality of motor skill learning in multi-goal and multi-constrained environments. An uncertainty-aware multi-timescale value function learning and selection method is proposed to overcome long horizon problems, improve optimal policy resolution and therefore enhance the generalization ability of goal-conditioned RL.

The purpose of this research is to provide insights into the daily search strategies of users, which can inform the enhancement of search experiences across multiple applications. By understanding how users navigate and interact with different apps during their search processes, the study seeks to contribute to the design of more intuitive and user-friendly app systems.

This study employs a mixed-methods approach to analyze users' daily search strategies in a natural cross-app interactive environment. Data collection was conducted using the Critical Incident Technique and the Micro-Moment Time Line, involving 204 participants to capture their real-time search experiences. Open coding techniques were utilized to categorize sequential search tactics, while the PrefixSpan algorithm was applied to identify patterns in frequently applied search strategies.

The study findings unveil a comprehensive framework that includes a variety of intra-app search tactics and inter-app switching tactics. Five predominant search strategies were identified: Iterative querying, Selective results adoption, Share-related, Recommended browsing, and Organizational results strategies. These strategies reflect the nuanced ways in which users engage with apps to fulfill their information needs.

This research represents a pioneering effort in systematically identifying and categorizing daily search strategies within a natural cross-app interaction context. It offers original contributions to the field by combining intra-app and inter-app tactics, providing a holistic view of user behavior. The implications of these findings are significant for app developers and designers, as they can leverage this knowledge to improve app functionality and user manuals, ultimately enhancing the overall search experience for users.

Employees facing challenges in their careers – i.e. female, migrant, elderly and lower-educated employees – might expect job searches to have a low likelihood of success and might therefore more often stay in unsatisfactory positions. The goal of this study is to discover inequalities in job mobility for these employees.

We rely on a large sample of Dutch public sector employees (N = 30,709) and study whether employees with challenges in their careers are hampered in translating job dissatisfaction into job searches. Additionally, we assess whether this is due to their perceptions of labor market alternatives.

Findings show that non-Western migrant, elderly and lower-educated employees are less likely to act on job dissatisfaction than their advantaged counterparts, whereas women are more likely than men to do so. Additionally, we find that although they perceive labor market opportunities as limited, this does not affect their propensity to search for different jobs.

This paper is novel in discovering inequalities in job mobility by analyzing whether employees facing challenges in their careers are less likely to act on job dissatisfaction and therefore more likely to remain in unsatisfactory positions.

In recent years, the demand for garments has significantly increased, requiring manufacturers to speed up their production to attract customers. Cut order planning (COP) is one of the most important processes in the apparel manufacturing industry. The appropriate stencil arrangement can reduce costs and fabric waste. The COP problem focuses on determining the size combination for a pattern, which is determined by the length of the cutting table, width, demand order, and height of the cutting equipment.

This study proposes new heuristics: genetic algorithm (GA), symbiotic organism search, and divide-and-search-based Lite heuristic and a One-by-One (ObO) heuristic to address the COP problem. The objective of the COP problem is to determine the optimal combination of stencils to meet demand requirements and minimize the total fabric length.

A comparison between our proposed heuristics and other simulated annealing and GA-based heuristics, and a hybrid approach (conventional algorithm + GA) was conducted to demonstrate the effectiveness and efficiency of the proposed heuristics. The test results show that the ObO heuristic can significantly improve the solution efficiency and find the near optimal solution for extreme demands.

This paper proposes a new heuristic, the One-by-One (ObO) heuristic, to solve the COP problem. The results show that the proposed approaches overcome the long operation time required to determine the fitting arrangement of stencils. In particular, our proposed ObO heuristic can significantly improve the solution efficiency, i.e. finding the near optimal solution for extreme demands within a very short time.

By introducing Stacking algorithm to solve the underfitting problem caused by insufficient data in traditional machine learning, this paper provides a new solution to the cold start problem of entrepreneurial borrowing risk control.

The authors introduce semi-supervised learning and integrated learning into the field of migration learning, and innovatively propose the Stacking model migration learning, which can independently train models on entrepreneurial borrowing credit data, and then use the migration strategy itself as the learning object, and use the Stacking algorithm to combine the prediction results of the source domain model and the target domain model.

The effectiveness of the two migration learning models is evaluated with real data from an entrepreneurial borrowing. The algorithmic performance of the Stacking-based model migration learning is further improved compared to the benchmark model without migration learning techniques, with the model area under curve value rising to 0.8. Comparing the two migration learning models reveals that the model-based migration learning approach performs better. The reason for this is that the sample-based migration learning approach only eliminates the noisy samples that are relatively less similar to the entrepreneurial borrowing data. However, the calculation of similarity and the weighing of similarity are subjective, and there is no unified judgment standard and operation method, so there is no guarantee that the retained traditional credit samples have the same sample distribution and feature structure as the entrepreneurial borrowing data.

From a practical standpoint, on the one hand, it provides a new solution to the cold start problem of entrepreneurial borrowing risk control. The small number of labeled high-quality samples cannot support the learning and deployment of big data risk control models, which is the cold start problem of the entrepreneurial borrowing risk control system. By extending the training sample set with auxiliary domain data through suitable migration learning methods, the prediction performance of the model can be improved to a certain extent and more generalized laws can be learned.

This paper introduces the thought method of migration learning to the entrepreneurial borrowing scenario, provides a new solution to the cold start problem of the entrepreneurial borrowing risk control system and verifies the feasibility and effectiveness of the migration learning method applied in the risk control field through empirical data.

The purpose of this paper is to systematically review the resilience literature to integrate the fragmented views and provide a more comprehensive understanding. This study aims to fill the gaps in the literature by discussing scientific contributions and highlighting the main issues in previous research findings regarding the definitions, dimensions and concepts that make up organizational resilience. This research highlights organizational resilience from a capabilities perspective.

This paper uses the systematic literature review method by searching all previous studies studying organizational resilience from 2014 to 2023. In total, there are 28 articles reviewed from the Scopus database.

This study found that resilience is a complex concept and has many definitions and dimensions. This study also conceptualizes organizational capability with a pyramid, which illustrates the basic framework of the six stages of the resilience process and hierarchically forms organizational resilience.

First, the keyword search strings on the repository database are currently limited to a few keywords. Need to broaden the range of keywords so as to produce a more comprehensive review. Second, the exclusion of books, book chapters and conference papers limits research findings and results. These sources are likely to enrich resilience development from various perspectives. Even though Scopus is the largest repository database, the research findings are not yet fully generalizable. Future researchers can add data from WoS, Ebsco or other databases. Literature obtained from various databases that can provide broader results.

The practical implications of this study are to provide a basis for managers in making decisions for organizational sustainability. Managers can consider each stage in the resilience capability pyramid as a reference for making strategic plans and relational orientation toward organizational members.

This research provides a hierarchical perspective on organizational resilience capabilities. For academics and practitioners, this study provides a critical and comprehensive systematization of the limited academic literature on resilience. This study also offers opportunities for further research to overcome the limitations of empirical testing of resilience capability construction using various theories and methodologies.

The purpose of this paper is to present an efficient path planning method for the multi-UAV system in target coverage problems.

An enhanced particle swarm optimizer (PSO) is used to solve the path planning problem, which concerns the two-dimensional motion of multirotor unmanned aerial vehicles (UAVs) in a three-dimensional environment. Enhancements include an improved initial swarm generation and prediction strategy for succeeding generations. Initial swarm improvements include the clustering process managed by fuzzy c-means clustering method, ordering procedure handled by ant colony optimizer and design vector change. Local solutions form the foundation of a prediction strategy.

Numerical simulations show that the proposed method could find near-optimal paths for multi-UAVs effectively.

Simulations indicate the proposed method could be deployed for autonomous multi-UAV systems with target coverage problems.

The proposed method combines intelligent methods in the early phase of PSO, handles obstacle avoidance problems with a unique approach and accelerates the process by adding a prediction strategy.