WZWei Zhang

Wei Zhang

*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


Wei Zhang has a total of 71 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    A Study on Methods for Evaluating the Quality of Datasets for Intelligent Algorithm Evaluation

    by Lu Bai, Yuqing Gu, Wei Zhang, Zilong Liu

    The credibility of artificial intelligence (AI) algorithm testing and evaluation directly determines the assessment of their intelligence level and application reliability, which heavily depends on the quality of input datasets. Currently, datasets for AI algorithm testing generally suffer from unclear sources, arbitrary annotations, and heterogeneous formats, while lacking quantitative evaluation specifications that integrate metrological characteristics with algorithm requirements, leading to poor comparability and low credibility of evaluation results. To address these gaps, this study proposes a dataset quality evaluation method oriented to AI algorithm testing by integrating metrological traceability and measurement uncertainty control concepts. Firstly, following the principles of scientificity, quantifiability, and traceability, a hierarchical technical framework and a full-process technical roadmap were constructed. Secondly, a five-dimensional quantitative evaluation index system was established, covering repeatability, completeness, accuracy, consistency, and traceability, with an accompanying uncertainty evaluation, where weighted scoring and measurement uncertainty evaluation methods were introduced to realize quantitative classification of dataset quality. Verification experiments on a temperature and humidity meter image dataset show that the proposed method can effectively assess and grade dataset quality levels. The evaluated dataset obtained a score of 0.9765 ± 0.010 (k = 2), reaching the “excellent” grade, demonstrating the framework’s ability to recognize high-quality datasets. This proof-of-concept demonstrates the framework’s feasibility for assessing high-quality datasets. Formal Standard Reference Data certification is beyond the scope of this study and requires further validation. This research provides a feasible path for the standardized construction of candidate reference data in the field of digital metrology and holds significant value for promoting the standardized application of AI technologies in critical domains such as measurement and control.

  • Open Access

    From Resource Access to Capability Conversion: Education, Employment, and Access to Smart-City Resources in Relation to Quality of Life Among B40 Households in Greater Klang Valley, Malaysia

    by Yifan Yang, Zhengyi Zhao, Lu Bai, Yanlin Liang, Paul Anthony Mariadas, Haiyang Jiang, Wei Zhang, Ruofan Zhao, Siti Nurul Munawwarah Roslan

    This study examined the associations among Education Affordability (EA), Job Opportunity (JO), Access to Smart-City Resources (ASR), and Quality of Life (QoL) among respondents from B40 households in Malaysia’s Greater Klang Valley. A cross-sectional, face-to-face survey was conducted in December 2024 using convenience sampling. Of 400 completed questionnaires, nine respondents did not meet the eligibility criteria, yielding a final analytic sample of 391. Confirmatory factor analysis assessed the four-factor measurement model, and PROCESS Model 59 estimated direct, indirect, and conditional associations using HC3 heteroscedasticity-consistent standard errors and 5000 bootstrap samples. EA was positively associated with JO and QoL, while JO and ASR were positively associated with QoL. ASR was not significantly associated with JO, and the EA × ASR interaction in the JO equation was not statistically significant. The indirect association between EA and QoL through JO was supported at low, mean, and high ASR levels, while pairwise contrasts were not statistically significant. In the QoL equation, the EA × ASR interaction was statistically significant, whereas the JO × ASR interaction was not. Income vulnerability differentiated the separate EA–QoL, JO–QoL, and ASR–QoL associations, while neither three-way interaction was statistically significant. Overall, the findings identify Education Affordability, employment opportunity, and resident-facing smart-city resources as distinct but interconnected correlates of Quality of Life among B40 households.

  • Open Access

    Assessing the Fitness-for-Purpose of Published Breath Analysis Data: A Quality Assessment Framework for Diabetes Biomarker Research

    by Lin Guo, Wei Zhang, Yinchu Wang, Fengyi Zhang, Zilong Liu, Xingchuang Xiong

    Published breath examination studies report inconsistent associations between diabetes and exhaled volatile organic compounds, partly because conventional appraisal tools do not assess whether quantitative data are comparable and fit for secondary use. We developed the National Institute of Metrology–Diabetes Breath Assessment (NIM-DBA) framework, drawing on metrological specifications and data-quality principles issued by the International Organization for Standardization (ISO). PubMed, Scopus, Embase, and Web of Science were searched through April 2025 for quantitative human breath data related to diabetes. Breath acetone was used as a case study, and the same literature pool was assessed with the Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) tool. Of 38 eligible studies, six (15.8%) met all NIM-DBA criteria. QUADAS-2 found high risk of bias in patient selection (79%) and unclear risk in index-test reporting (82%). A post-classification comparison of the six criteria-met and 32 criteria-not-met studies identified exploratory enrichment of standardized breath maneuvers, calibration coverage, alveolar or end-expiratory sampling, pre-sampling alcohol control, sample-stability validation, background assessment, and morning sampling. These findings do not establish breath acetone as a standalone diagnostic biomarker. Instead, they show how fitness-for-purpose assessment can identify more comparable evidence and clarify how methodological choices contribute to inconsistent findings. NIM-DBA complements clinical risk-of-bias tools and may support secondary quantitative data use and future study design.

  • Open Access

    MSIMG: A Density-Aware Multi-Channel Image Representation Method for Mass Spectrometry

    by Fengyi Zhang, Boyong Gao, Yinchu Wang, Lin Guo, Wei Zhang, Xingchuang Xiong

    Extracting key features for phenotype classification from high-dimensional and complex mass spectrometry (MS) data presents a significant challenge. Conventional data representation methods, such as traditional peak lists or grid-based imaging strategies, are often hampered by information loss and compromised signal integrity, thereby limiting the performance of downstream deep learning models. To address this issue, we propose a novel data representation framework named MSIMG. Inspired by object detection in computer vision, MSIMG introduces a data-driven, “density-peak-centric” patch selection strategy. This strategy employs density map estimation and non-maximum suppression algorithms to locate the centers of signal-dense regions, which serve as anchors for dynamic, content-aware patch extraction. This process transforms raw mass spectrometry data into a multi-channel image representation with higher information fidelity. Extensive experiments conducted on two public clinical mass spectrometry datasets demonstrate that MSIMG significantly outperforms both the traditional peak list method and the grid-based MetImage approach. This study confirms that the MSIMG framework, through its content-aware patch selection, provides a more information-dense and discriminative data representation paradigm for deep learning models. Our findings highlight the decisive impact of data representation on model performance and successfully demonstrate the immense potential of applying computer vision strategies to analytical chemistry data, paving the way for the development of more robust and precise clinical diagnostic models.


  • Coupling Supercritical Fluid Chromatography with Internal Standard Corrected Quantitative NMR for Rapid and Accurate Purity Assessment

    by Huan Yao, Ting Huang, Zhifang Li, Huaxin Wu, Xueyao Li, Hua Tang, Fuhai Su, Xinliang Jin, Wei Zhang

    Purity assessment is critical for the quality control of active pharmaceutical ingredients (APIs). However, there is a lack of rapid and accurate purity assessment methods that operate independently of target-specific certified reference materials (CRMs). Herein, we proposed a traceable quantification method, named SFC-ISC-qNMR, for purity assessment and quantitative determination of low-purity APIs, which was accomplished by tracing the test solution to the reference solution via supercritical fluid chromatography (SFC) collection with internal standard-correction (ISC), followed by an SI-traceable quantitative nuclear magnetic resonance (qNMR) analysis of the reference solution (SI = International System of Units). Digitoxin, a widely used cardiac glycoside in the treatment of heart failure and arrhythmias, was used as a model compound. By employing a mobile phase of supercritical CO2 modified with 24%-26% methanol, digitoxin with a purity exceeding 99.87% was isolated and collected. A national CRM of ethyl para-hydroxybenzoate was used as the IS for digitoxin. The SFC-ISC reference solution can be prepared in 15 min without the need for precise weighing, simply by reconstitution of the SFC collected high-purity analyte in deuterated methanol containing IS. The feasibility and reliability of the SFC-ISC-qNMR method were exhibited through the quantitative determination of digitoxin in chemical mixtures. The purity of a commercially available digitoxin reagent was determined to be 94.17% ± 0.15%, which was consistent with high-resolution qNMR analysis. The contents of spiked digitoxin in a digoxin analytical standard were measured at 1.06%, 1.37%, 2.39%, and 4.43%, respectively, showing high consistency with the gravimetric value with a bias less than 0.13%. In contrast, direct qNMR with deconvolution and HPLC analysis failed to provide accurate measurements. These results highlight the applicability of the SFC-ISC-qNMR method for drug evaluation and research, demonstrating its significant potential for the purity assessment of APIs.


  • Open Access

    EQLC-EC: An Efficient Voting Classifier for 1D Mass Spectrometry Data Classification

    by Lin Guo, Yinchu Wang, Zilong Liu, Fengyi Zhang, Wei Zhang, Xingchuang Xiong

    Mass spectrometry (MS) data present challenges for machine learning (ML) classification due to their high dimensionality, complex feature distributions, batch effects, and intensity discrepancies, often hindering model generalization and efficiency. To address these issues, this study introduces the Efficient Quick 1D Lite Convolutional Neural Network (CNN) Ensemble Classifier (EQLC-EC), integrating 1D convolutional networks with reshape layers and dual voting mechanisms for enhanced feature representation and classification performance. Validation was performed on five publicly available MS datasets, each featured in high-impact publications. EQLC-EC underwent comprehensive evaluation against classical machine learning (ML) models (e.g., support vector machine (SVM), random forest) and the leading deep learning methods reported in these studies. EQLC-EC demonstrated dataset-specific improvements, including enhanced classification accuracy (1–5% increase) and reduced standard deviation (1–10% reduction). Performance differences between soft and hard voting mechanisms were negligible (<1% variation in accuracy and standard deviation). EQLC-EC presents a powerful and efficient tool for MS data analysis with potential applications across metabolomics and proteomics.


  • Post-collection purity correction for internal standard correction-high performance liquid chromatography-quantitative nuclear magnetic resonance

    by Xueyao Li, Wei Zhang, Ting Huang, Ming Li, Fuhai Su, Huaxin Wu, Guangshi Tang

    The improved method enhances the accuracy of quantifying low-purity organic compounds by analyzing eluate components for correction factors, validated through oxytetracycline quantification, addressing qNMR challenges and ensuring purity.


  • Accurate Quantification of Pure Thiacloprid with Mass Balance and Quantitative H-NMR

    by Mengling Tu, Wei Zhang, Yunxiao Zhu, Wen Ma, Xiao Li, Shukun Zhou, Hongmei Li, Xianjiang Li


  • Accurate Determination of Trace Water in Organic Solution by Quantitative Nuclear Magnetic Resonance

    by Kangni Wan, Ming Li, Ting Huang, Wei Zhang, Tianji Zhang, Xiuqin Li, Haifeng Wang, Juan Lv

    A new method accompanied by a derived equation for accurate determination of trace water was developed by using quantitative 1H nuclear magnetic resonance (qNMR) spectroscopy. Given that the response for each chemically distinct moiety is uniformly proportional to the number of the corresponding resonant nuclei within the analyte, it is practicable to directly quantify the water content via its proton number using qNMR. In this study, three water standards with known water contents (e.g., 10.02, 1.006, and 0.103 mg/g), which were accurately determined by a well-established Coulometric Karl Fischer (CKF) titration method, were measured by using the developed qNMR method. An excellent agreement between the results from these two methods was obtained. Then, the water content of Sudan I was determined by high-field NMR (HF-NMR) spectroscopy, and the water contents of acetone and bioethanol were measured by low-field NMR (LF-NMR) spectroscopy. These results were compared with the water content measured by the CKF method to confirm the applicability of the established qNMR method. The developed method can eliminate the influences of environmental humidity and background water in the solvent; subsequently, the results calculated by the derived equation were comparable to the nominal values. Under the optimal conditions, the limit of quantitation of this method was as low as 6.7 μg. The recommended sample sizes for practical samples with various water contents (e.g., 10.02, 1.006, and 0.103 mg/g) were determined to be 5, 50, and 60 mg, respectively, which are much smaller than those required for the CKF method. The new method has a static and stable process without any side reactions, and the traceability to the SI unit can be directly achieved through the NMR internal standard. This method overcomes the limitations of the CKF method, especially for measuring methanol-insoluble substances.


  • Open Access

    Purity Assessment of Dinotefuran Using Mass Balance and Quantitative Nuclear Magnetic Resonance

    by Xianjiang Li, Wei Zhang, Xiao Li, Shukun Zhou, Mengling Tu, Yunxiao Zhu, Hongmei Li

    Dinotefuran (DNT) belongs to the third-generation neonicotinoid pesticides, which are among the most common residuals in a variety of food commodities. To guarantee accurate quantification and traceability of results in food samples, certified reference materials (CRMs) are the indispensable benchmark. In this work, a DNT CRM was characterized and its purity was assessed by two independent methods, including mass balance (MB) and quantitative nuclear magnetic resonance spectroscopy (qNMR). The mass fraction of moisture was 0.33 mg/g, the inorganic impurity was 0.01 mg/g, and no detectable organic solvent was detected. Benzoic acid was chosen as the internal standard for qNMR. Its mass fraction was 997.9 mg/g and 992.9 mg/g by MB and qNMR, respectively. Eventually, the DNT CRM was assigned a mass fraction of 995 mg/g, with expanded uncertainty of 5 mg/g (k = 2). This CRM can be used to prepare calibrant solutions and is applicable to national routine monitoring of DNT residuals in agro-products and food.