Xu Wang
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
Xu Wang has a total of 18 co-authors. Here are the top co-authors by number of shared publications:
Kun Zhu
Wuhan Univ, Sch Comp Sci, Wuhan 430072, Peoples R China
4 shared publicationsNana Zhang
Wuhan Univ, Sch Comp Sci, Wuhan 430072, Peoples R China
4 shared publications- Shi Ying
School of Computer Science Wuhan University No.299 Bayi Road Wuhan People's Republic of China
2 shared publications - Christopher Brewster
Department of Advanced Computing Sciences, Maastricht University, 6200 MD Maastricht, The Netherlands
1 shared publication Feijing Liu
Shanghai University
1 shared publication- Frank van Harmelen
Vrije Universiteit Amsterdam
1 shared publication
Latest Publications
- Open Access
- 18 April 2025
- Published by MDPI AG in Proceedings
- Vol. 117 (1) , 5
Unveiling the Secrets of Embeddings: Does the Importance of Agricultural Terms Relate to the Context They Occur In?
Advances in language modeling have provided affordances for term extraction based on the capture of the lexical context and its semantics and encoding this in real-valued vectors (embeddings). Term importance is usually computed using quantitative measures, ignoring the semantic context. Until now, there has been limited or no research on the effect of context on term importance using machine learning methods. In this paper, we investigate whether there is a relation between the importance of agricultural terms and the context of their occurrence as represented by text embeddings. Using a dataset of almost 33.7 K AGRIS abstracts containing 50 concepts randomly extracted from AGROVOC, we computed the correlation between the concept tf-idf scores and each of three semantic distances (cosine similarity of embeddings) used as a proxy to context: (i) the semantic distances of the various occurrences of the concept; (ii) the semantic distances between the concept and the abstracts that it appears in; (iii) the semantic distances of the abstracts in which the concept occurs. Embeddings were generated using Agriculture-BERT. We present a methodology and initial results from the computation of correlations. The novelty of our work is in the systematic investigation of the relation between term importance and semantic context.
- 18 April 2025
- Published by MDPI AG in Proceedings
- Vol. 117 (1) , 5
Open AccessUnveiling the Secrets of Embeddings: Does the Importance of Agricultural Terms Relate to the Context They Occur In?
Advances in language modeling have provided affordances for term extraction based on the capture of the lexical context and its semantics and encoding this in real-valued vectors (embeddings). Term importance is usually computed using quantitative measures, ignoring the semantic context. Until now, there has been limited or no research on the effect of context on term importance using machine learning methods. In this paper, we investigate whether there is a relation between the importance of agricultural terms and the context of their occurrence as represented by text embeddings. Using a dataset of almost 33.7 K AGRIS abstracts containing 50 concepts randomly extracted from AGROVOC, we computed the correlation between the concept tf-idf scores and each of three semantic distances (cosine similarity of embeddings) used as a proxy to context: (i) the semantic distances of the various occurrences of the concept; (ii) the semantic distances between the concept and the abstracts that it appears in; (iii) the semantic distances of the abstracts in which the concept occurs. Embeddings were generated using Agriculture-BERT. We present a methodology and initial results from the computation of correlations. The novelty of our work is in the systematic investigation of the relation between term importance and semantic context.
- Open Access
- 8 June 2022
- Published by SAGE Publications in Data Science
- Vol. 5 (2) , 167-193
Recommending scientific datasets using author networks in ensemble methods
Open access to datasets is increasingly driving modern science. Consequently, discovering such datasets is becoming an important functionality for scientists in many different fields. We investigate methods for dataset recommendation: the task of recommending relevant datasets given a dataset that is already known to be relevant. Previous work has used meta-data descriptions of datasets and interest profiles of authors to support dataset recommendation. In this work, we are the first to investigate the use of co-author networks to drive the recommendation of relevant datasets. We also investigate the combination of such co-author networks with existing methods, resulting in three different algorithms for dataset recommendation. We obtain experimental results on a realistic corpus which show that only the ensemble combination of all three algorithms achieves sufficiently high precision for the dataset recommendation task.
- 8 June 2022
- Published by SAGE Publications in Data Science
- Vol. 5 (2) , 167-193
Open AccessRecommending scientific datasets using author networks in ensemble methods
Open access to datasets is increasingly driving modern science. Consequently, discovering such datasets is becoming an important functionality for scientists in many different fields. We investigate methods for dataset recommendation: the task of recommending relevant datasets given a dataset that is already known to be relevant. Previous work has used meta-data descriptions of datasets and interest profiles of authors to support dataset recommendation. In this work, we are the first to investigate the use of co-author networks to drive the recommendation of relevant datasets. We also investigate the combination of such co-author networks with existing methods, resulting in three different algorithms for dataset recommendation. We obtain experimental results on a realistic corpus which show that only the ensemble combination of all three algorithms achieves sufficiently high precision for the dataset recommendation task.
- Open Access
- 1 January 2020
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 65 (2) , 1467-1486
Software defect prediction based on non-linear manifold learning and hybrid deep learning techniques
Software defect prediction plays a very important role in software quality assurance, which aims to inspect as many potentially defect-prone software modules as possible. However, the performance of the prediction model is susceptible to high dimensionality of the dataset that contains irrelevant and redundant features. In addition, software metrics for software defect prediction are almost entirely traditional features compared to the deep semantic feature representation from deep learning techniques. To address these two issues, we propose the following two solutions in this paper: (1) We leverage a novel non-linear manifold learning method - SOINN Landmark Isomap (SLIsomap) to extract the representative features by selecting automatically the reasonable number and position of landmarks, which can reveal the complex intrinsic structure hidden behind the defect data. (2) We propose a novel defect prediction model named DLDD based on hybrid deep learning techniques, which leverages denoising autoencoder to learn true input features that are not contaminated by noise, and utilizes deep neural network to learn the abstract deep semantic features. We combine the squared error loss function of denoising autoencoder with the cross entropy loss function of deep neural network to achieve the best prediction performance by adjusting a hyperparameter. We compare the SL-Isomap with seven state-of-the-art feature extraction methods and compare the DLDD model with six baseline models across 20 open source software projects. The experimental results verify that the superiority of SL-Isomap and DLDD on four evaluation indicators.
- 1 January 2020
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 65 (2) , 1467-1486
Open AccessSoftware defect prediction based on non-linear manifold learning and hybrid deep learning techniques
Software defect prediction plays a very important role in software quality assurance, which aims to inspect as many potentially defect-prone software modules as possible. However, the performance of the prediction model is susceptible to high dimensionality of the dataset that contains irrelevant and redundant features. In addition, software metrics for software defect prediction are almost entirely traditional features compared to the deep semantic feature representation from deep learning techniques. To address these two issues, we propose the following two solutions in this paper: (1) We leverage a novel non-linear manifold learning method - SOINN Landmark Isomap (SLIsomap) to extract the representative features by selecting automatically the reasonable number and position of landmarks, which can reveal the complex intrinsic structure hidden behind the defect data. (2) We propose a novel defect prediction model named DLDD based on hybrid deep learning techniques, which leverages denoising autoencoder to learn true input features that are not contaminated by noise, and utilizes deep neural network to learn the abstract deep semantic features. We combine the squared error loss function of denoising autoencoder with the cross entropy loss function of deep neural network to achieve the best prediction performance by adjusting a hyperparameter. We compare the SL-Isomap with seven state-of-the-art feature extraction methods and compare the DLDD model with six baseline models across 20 open source software projects. The experimental results verify that the superiority of SL-Isomap and DLDD on four evaluation indicators.
- Open Access
- 1 January 2020
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 65 (1) , 279-308
Software Defect Prediction Based on Stacked Contractive Autoencoder and Multi-objective Optimization
Software defect prediction plays an important role in software quality assurance. However, the performance of the prediction model is susceptible to the irrelevant and redundant features. In addition, previous studies mostly regard software defect prediction as a single objective optimization problem, and multi-objective software defect prediction has not been thoroughly investigated. For the above two reasons, we propose the following solutions in this paper: (1) we leverage an advanced deep neural network—Stacked Contractive AutoEncoder (SCAE) to extract the robust deep semantic features from the original defect features, which has stronger discrimination capacity for different classes (defective or non-defective). (2) we propose a novel multi-objective defect prediction model named SMONGE that utilizes the Multi-Objective NSGAII algorithm to optimize the advanced neural network—Extreme learning machine (ELM) based on state-of-the-art Pareto optimal solutions according to the features extracted by SCAE. We mainly consider two objectives. One objective is to maximize the performance of ELM, which refers to the benefit of the SMONGE model. Another objective is to minimize the output weight norm of ELM, which is related to the cost of the SMONGE model. We compare the SCAE with six state-of-the-art feature extraction methods and compare the SMONGE model with multiple baseline models that contain four classic defect predictors and the MONGE model without SCAE across 20 open source software projects. The experimental results verify that the superiority of SCAE and SMONGE on seven evaluation metrics.
- 1 January 2020
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 65 (1) , 279-308
Open AccessSoftware Defect Prediction Based on Stacked Contractive Autoencoder and Multi-objective Optimization
Software defect prediction plays an important role in software quality assurance. However, the performance of the prediction model is susceptible to the irrelevant and redundant features. In addition, previous studies mostly regard software defect prediction as a single objective optimization problem, and multi-objective software defect prediction has not been thoroughly investigated. For the above two reasons, we propose the following solutions in this paper: (1) we leverage an advanced deep neural network—Stacked Contractive AutoEncoder (SCAE) to extract the robust deep semantic features from the original defect features, which has stronger discrimination capacity for different classes (defective or non-defective). (2) we propose a novel multi-objective defect prediction model named SMONGE that utilizes the Multi-Objective NSGAII algorithm to optimize the advanced neural network—Extreme learning machine (ELM) based on state-of-the-art Pareto optimal solutions according to the features extracted by SCAE. We mainly consider two objectives. One objective is to maximize the performance of ELM, which refers to the benefit of the SMONGE model. Another objective is to minimize the output weight norm of ELM, which is related to the cost of the SMONGE model. We compare the SCAE with six state-of-the-art feature extraction methods and compare the SMONGE model with multiple baseline models that contain four classic defect predictors and the MONGE model without SCAE across 20 open source software projects. The experimental results verify that the superiority of SCAE and SMONGE on seven evaluation metrics.
- 1 January 2020
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 64 (1) , 471-499
KAEA: A Novel Three-stage Ensemble Model for Software Defect Prediction
Software defect prediction is a research hotspot in the field of software engineering. However, due to the limitations of current machine learning algorithms, we can’t achieve good effect for defect prediction by only using machine learning algorithms. In previous studies, some researchers used extreme learning machine (ELM) to conduct defect prediction. However, the initial weights and biases of the ELM are determined randomly, which reduces the prediction performance of ELM. Motivated by the idea of search based software engineering, we propose a novel software defect prediction model named KAEA based on kernel principal component analysis (KPCA), adaptive genetic algorithm, extreme learning machine and Adaboost algorithm, which has three main advantages: (1) KPCA can extract optimal representative features by leveraging a nonlinear mapping function; (2) We leverage adaptive genetic algorithm to optimize the initial weights and biases of ELM, so as to improve the generalization ability and prediction capacity of ELM; (3) We use the Adaboost algorithm to integrate multiple ELM basic predictors optimized by adaptive genetic algorithm into a strong predictor, which can further improve the effect of defect prediction. To effectively evaluate the performance of KAEA, we use eleven datasets from large open source projects, and compare the KAEA with four machine learning basic classifiers, ELM and its three variants. The experimental results show that KAEA is superior to these baseline models in most cases.
- 1 January 2020
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 64 (1) , 471-499
KAEA: A Novel Three-stage Ensemble Model for Software Defect Prediction
Software defect prediction is a research hotspot in the field of software engineering. However, due to the limitations of current machine learning algorithms, we can’t achieve good effect for defect prediction by only using machine learning algorithms. In previous studies, some researchers used extreme learning machine (ELM) to conduct defect prediction. However, the initial weights and biases of the ELM are determined randomly, which reduces the prediction performance of ELM. Motivated by the idea of search based software engineering, we propose a novel software defect prediction model named KAEA based on kernel principal component analysis (KPCA), adaptive genetic algorithm, extreme learning machine and Adaboost algorithm, which has three main advantages: (1) KPCA can extract optimal representative features by leveraging a nonlinear mapping function; (2) We leverage adaptive genetic algorithm to optimize the initial weights and biases of ELM, so as to improve the generalization ability and prediction capacity of ELM; (3) We use the Adaboost algorithm to integrate multiple ELM basic predictors optimized by adaptive genetic algorithm into a strong predictor, which can further improve the effect of defect prediction. To effectively evaluate the performance of KAEA, we use eleven datasets from large open source projects, and compare the KAEA with four machine learning basic classifiers, ELM and its three variants. The experimental results show that KAEA is superior to these baseline models in most cases.
- 1 January 2020
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 62 (3) , 1161-1185
A Performance Fault Diagnosis Method for SaaS Software Based on GBDT Algorithm
SaaS software that provides services through cloud platform has been more widely used nowadays. However, when SaaS software is running, it will suffer from performance fault due to factors such as the software structural design or complex environments. It is a major challenge that how to diagnose software quickly and accurately when the performance fault occurs. For this challenge, we propose a novel performance fault diagnosis method for SaaS software based on GBDT (Gradient Boosting Decision Tree) algorithm. In particular, we leverage the monitoring mean to obtain the performance log and warning log when the SaaS software system runs, and establish the performance fault type set and determine performance log feature. We also perform performance fault type annotation for the performance log combined with the analysis result of the warning log. Moreover, we deal with the incomplete performance log and the type non-equalization problem by using the mean filling for the same type and combination of SMOTE (Synthetic Minority Oversampling Technique) and undersampling methods. Finally, we conduct an empirical study combined with the disaster reduction system deployed on the cloud platform, and it demonstrates that the proposed method has high efficiency and accuracy for the performance diagnosis when SaaS software system runs.
- 1 January 2020
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 62 (3) , 1161-1185
A Performance Fault Diagnosis Method for SaaS Software Based on GBDT Algorithm
SaaS software that provides services through cloud platform has been more widely used nowadays. However, when SaaS software is running, it will suffer from performance fault due to factors such as the software structural design or complex environments. It is a major challenge that how to diagnose software quickly and accurately when the performance fault occurs. For this challenge, we propose a novel performance fault diagnosis method for SaaS software based on GBDT (Gradient Boosting Decision Tree) algorithm. In particular, we leverage the monitoring mean to obtain the performance log and warning log when the SaaS software system runs, and establish the performance fault type set and determine performance log feature. We also perform performance fault type annotation for the performance log combined with the analysis result of the warning log. Moreover, we deal with the incomplete performance log and the type non-equalization problem by using the mean filling for the same type and combination of SMOTE (Synthetic Minority Oversampling Technique) and undersampling methods. Finally, we conduct an empirical study combined with the disaster reduction system deployed on the cloud platform, and it demonstrates that the proposed method has high efficiency and accuracy for the performance diagnosis when SaaS software system runs.
- 20 April 2016
- Published by Springer Nature
- p. 561-573
A Method of Event Ontology Mapping