SCShenglei Chen

Shenglei Chen

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Shenglei Chen has a total of 18 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Cross-Entropy-Weighted Model-Selection Averaged One-Dependence Estimators

    by Shenglei Chen, Chengzhen Zhang

    Model selection has been applied to AODE (Averaged One-Dependence Estimator) algorithms in previous studies, significantly improving its classification performance. This is particularly effective for high-dimensional and complex datasets, which improves both the accuracy and robustness of the algorithm. Nevertheless, current cross-validation-based risk minimization strategies are still relatively limited in their application to AODE algorithms. In addition, these strategies usually do not fully account for the differences between different attributes, so the model does not accurately capture the associations and differences between attributes, which has an impact on the overall classification effect. To this end, we propose a Cross-Entropy-Weighted Model-Selection AODE (ESAODE) algorithm, which introduces a novel weight-computing scheme: epsilon-smoothed reciprocal cross entropy. It assigns distinct weights to all ODE sub-models according to the cross entropy obtained when each One-Dependence Estimator (ODE) approximates the target distribution. Furthermore, we evaluate the weighted sub-models and determine the optimal one based on empirical metrics from leave-one-out cross-validation (LOOCV). The new algorithm combines cross-entropy-based weighting and model selection within the AODE framework. Experimental results indicate that the ESAODE algorithm effectively integrates the advantages of the weighting mechanism and model selection, and improves the ability to capture key features as well as the adaptability to different datasets.

  • Attribute Value Weighted Averaged One-Dependence Estimators with Kullback–Leibler Divergence

    by Changjian Zhu, Shenglei Chen, Huihang Ke, Chengzhen Zhang

    The Averaged One-Dependence Estimators (AODE) algorithm is an improvement of the naive Bayes algorithm, which allows all the attributes dependent on one common attribute, called parent attribute, thus forming One-Dependence Estimators (ODE). The classification probability is estimated by averaging the conditional probability of the ODE. When there is a dependency relationship between attributes, the AODE algorithm can better capture these relationships, thus improving classification performance. The AODE algorithm treats the parent and child attribute values equally in different ODEs. However, the parent attribute value and the child attribute value in different ODEs have different importance for classification. In this paper, two attribute value weighted AODE based on Kullback–Leibler divergence are proposed, one is parent attribute value weighted AODE, and the other is child attribute value weighted AODE. Comparative experiments were carried out with 30 datasets in the UCI database, and the experiments indicate that the performance of the parent attribute value weighted AODE algorithm with Kullback–Leibler divergence is significantly better than the original AODE algorithm, and its performance is also better than the mutual information weighted AODE algorithm.


  • Open Access

    Research on Model Selection-Based Weighted Averaged One-Dependence Estimators

    by Chengzhen Zhang, Shenglei Chen, Huihang Ke

    The Averaged One-Dependence Estimators (AODE) is a popular and effective method of Bayesian classification. In AODE, selecting the optimal sub-model based on a cross-validated risk minimization strategy can further enhance classification performance. However, existing cross-validation risk minimization strategies do not consider the differences in attributes in classification decisions. Consequently, this paper introduces an algorithm for Model Selection-based Weighted AODE (SWAODE). To express the differences in attributes in classification decisions, the ODE corresponding to attributes are weighted, with mutual information commonly used in the field of machine learning adopted as weights. Then, these weighted sub-models are evaluated and selected using leave-one-out cross-validation (LOOCV) to determine the best model. The new method can improve the accuracy and robustness of the model and better adapt to different data features, thereby enhancing the performance of the classification algorithm. Experimental results indicate that the algorithm merges the benefits of weighting with model selection, markedly enhancing the classification efficiency of the AODE algorithm.


  • Open Access

    Balancing Privacy Risk and Benefit in Service Selection for Multiprovision Cloud Service Composition

    by Linyuan Liu, Haibin Zhu, Shenglei Chen

    The popularity of cloud computing has fueled the growth in multiprovision cloud service composition (MPCSC), where each cloud service provider (CSP) can fulfill multiple tasks, i.e., offer multiple services, simultaneously. In the MPCSC, users would rather disclose some private data for more benefits (e.g., personalized services). However, the more private data is released, the more serious the privacy risk faced by users. In particular, the multiservice provision characteristic of MPCSC further exacerbates the privacy risk. Therefore, how to balance the privacy risk and benefit in service selection for MPCSC is a challenging research problem. In this paper, firstly we explore the service selection problem of balancing privacy risk and benefit in MPCSC (SSBM), then we propose an improved Kuhn–Munkres (KM) algorithm solution to the SSBM problem. Furthermore, we conduct a series of simulation experiments to evaluate the proposed approach. The experimental results show that the proposed approach is both efficient and effective for solving the SSBM problem.

  • Selective AnDE based on attributes ranking by Maximin Conditional Mutual Information (MMCMI)

    by Shenglei Chen, Xin Ma, Linyuan Liu, Limin Wang

    Attribute selection has been proved to be an effective trick to strengthen the classification capability of Bayesian network classifiers, such as Averaged n-Dependence Estimators (AnDE). However, conventional mutual information-based attribute ranking considers only the correlation between the attribute and the class, regardless of the redundancies among the attributes. In this paper, we propose a new ranking approach, called Maximin Conditional Mutual Information (MMCMI), which first minimises the conditional mutual information for any unsorted attribute with regard to the sorted attribute sequence, and then maximise the minimal conditional mutual information within all unsorted attributes. When ranking the very first attribute, the mutual information with the class is maximised within all attributes. Extensive empirical results demonstrate that the MMCMI ranking approach together with attribute selection framework achieves significantly superior classification performance and less classification time with respect to regular AnDE and the mutual information counterparts.


  • Open Access

    Attribute Selecting in Tree-Augmented Naive Bayes by Cross Validation Risk Minimization

    by Shenglei Chen, Zhonghui Zhang, Linyuan Liu

    As an important improvement to naive Bayes, Tree-Augmented Naive Bayes (TAN) exhibits excellent classification performance and efficiency since it allows that every attribute depends on at most one other attribute in addition to the class variable. However, its performance might be lowered as some attributes might be redundant. In this paper, we propose an attribute Selective Tree-Augmented Naive Bayes (STAN) algorithm which builds a sequence of approximate models each involving only the top certain attributes and searches the model to minimize the cross validation risk. Five different approaches to ranking the attributes have been explored. As the models can be evaluated simultaneously in one pass learning through the data, it is efficient and can avoid local optima in the model space. The extensive experiments on 70 UCI data sets demonstrated that STAN achieves superior performance while maintaining the efficiency and simplicity.


  • Sample-Based Attribute Selective A$n$ DE for Large Data

    by Shenglei Chen, Ana M. Martinez, Geoffrey I. Webb, Limin Wang

    More and more applications have come with large data sets in the past decade. However, existing algorithms cannot guarantee to scale well on large data. Averaged n-Dependence Estimators (AnDE) allows for flexible learning from out-of-core data, by varying the value of n (number of super parents). Hence, AnDE is especially appropriate for large data learning. In this paper, we propose a sample-based attribute selection technique for AnDE. It needs one more pass through the training data, in which a multitude of approximate AnDE models are built and efficiently assessed by leave-one-out cross validation. The use of a sample reduces the training time. Experiments on 15 large data sets demonstrate that the proposed technique significantly reduces AnDE's error at the cost of a modest increase in training time. This efficient and scalable out-of-core approach delivers superior or comparable performance to typical in-core Bayesian network classifiers.


  • Selective AnDE for large data learning: a low-bias memory constrained approach

    by Shenglei Chen, Ana M. Martínez, Geoffrey I. Webb, Limin Wang


  • Multi-Step Truncated Q Learning Algorithm

    by Sheng-Lei Chen, Hui-Zhong Wu, Xiang-Lan Han, Liang Xiao

    Q learning is of great importance in reinforcement learning. To compensate the drawbacks of Q learning and Q(λ) algorithm, MTQ algorithm is proposed in this paper. It makes use of future information of k steps to update current Q value. Thus it can consider more long-term benefit and the computation complexity is also decreased. Good balance is achieved between update speed and computation complexity. Experiments demonstrate effectiveness of this algorithm.