The rapid growth of big data and complex machine learning models—gradient-boosted trees and deep neural networks—has produced highly accurate but opaque"black-box"predictors across medicine, finance, and industry, making interpretability a central concern in data analysis. SHapley Additive exPlanations (SHAP), grounded in cooperative game theory, has become one of the most influential interpretability methods because it provides theoretically consistent feature attributions at both the local and global levels. This paper presents a systematic literature review of SHAP and its role in data analysis. It synthesizes SHAP's theoretical foundations, its main implementations (TreeSHAP, KernelSHAP, and DeepSHAP), its visualization toolkit, and its practical applications, and it reports a compact empirical study comparing the three explainers on a clinical dataset. This study finds that, although SHAP markedly improves transparency and decision support, open challenges remain in computational cost, the reliability of explanations under feature correlation, and consistency across methods. The significance of this work is twofold: theoretically, it organizes SHAP's variants and properties within a single coherent framework; practically, it offers data analysts a structured, evidence-based reference for selecting and applying SHAP appropriately, thereby supporting more transparent, reliable, and accountable model-driven decisions.

Parametric models widely used in estimation often violate assumptions in practical research. As flexible alternatives, machine learning methods, especially tree ensembles and deep neural networks, impose fewer prior assumptions on functional forms for parameter estimation. This paper systematically examines eight estimation methods—ordinary least squares, ridge regression, lasso, random forest, XGBoost, LightGBM, multilayer perceptron (MLP), and deep neural networks—across four simulation regimes (linear, semiparametric, nonlinear, and high-dimensional sparse) and two real-world datasets: the Home Credit Default Risk dataset (307,511 samples) and the PIMA Indians Diabetes dataset (768 samples). This study evaluate model bias, mean squared error (MSE) and computational overhead to determine the applicable scenarios for each method category. Experimental results demonstrate that tree-based methods perform steadily across various scenarios. Although deep neural networks incur higher computational overhead, they achieve the minimal MSE when facing strong nonlinearity with moderate or large sample sizes. These findings provide actionable guidance for selecting estimation methods based on data characteristics, bridging theoretical advances in machine learning and practical estimation. The results verify that no single estimator outperforms all others across all data scenarios. The optimal selection relies on the joint effects of data nonlinearity, dimensionality and sample size, which highlights the necessity of diagnosis-oriented method selection in empirical studies.

Artificial intelligence (AI) in breast imaging has progressed from retrospective algorithm testing to prospective and real-world evaluation. In full-field digital mammography (FFDM), randomized, prospective, and implementation studies show that AI-supported reading maintains or increases cancer detection while reducing radiologist workload when thresholds, arbitration rules, and audit procedures are prespecified. Evidence for digital breast tomosynthesis (DBT) is expanding, but remains weighted toward reader studies and retrospective external validation rather than population-level deployment. Mammography-derived risk models extend AI from lesion detection to longitudinal risk stratification, provided that absolute-risk calibration and subgroup monitoring are performed locally. In breast MRI, AI triage, lesion classification, ultrafast acquisition, and reconstruction methods address access and workload constraints, although safety depends on conservative operating points and interval cancer surveillance. This paper draws the conclusion with a five-layer translation framework linking technical validity, workflow fit, safety monitoring, fairness, and economic sustainability, which offers references for further research in the applications of artificial intelligence in breast cancer imaging.

The rapid growth of online shopping platforms has significantly increased the analytical value of product review data. Short reviews on Chinese e-commerce platforms, typically containing fewer than 50 characters, simultaneously address multiple dimensions, including product quality, logistics, customer service, and value for money, rendering conventional sentence-level sentiment classification insufficient for fine-grained analysis. This study adopts aspect-based sentiment analysis (ABSA) as the core framework and constructs a labeling schema comprising four sentiment dimensions and three polarity levels, yielding 12 classification categories. Rule-based automated annotation was applied to 62,770 publicly available Chinese consumer reviews; manual verification on a 500-sample subset confirmed an annotation accuracy of 87.2%. Three lightweight models—TextCNN, BiLSTM with attention (BiLSTM+Attn), and ERNIE 3.0-Nano—were trained and evaluated under identical experimental conditions on an NVIDIA RTX 4060 laptop GPU. ERNIE 3.0-Nano proved to be the most accurate, with an accuracy of 0.6363 and a macro-average F1 score of 0.5798. BiLSTM+Attn was the fastest on GPU: 0.004 ms per sample. TextCNN had the smallest parameter count: 0.69 million. These findings provide quantitative evidence for model selection under diverse deployment constraints.

With the rapid deployment of neural machine translation (NMT) and large language models (LLMs), AI-assisted translation has become a cornerstone of multilingual communication. Despite achieving impressive fluency, these systems often perpetuate subtle yet systematic cultural biases embedded in training corpora, model architectures, and inference pipelines. This paper presents a systematic review of cultural bias in AI translation, organized around three research questions: (1) how cultural bias manifests, (2) how it can be identified, and (3) how it can be mitigated. Drawing on recent advances in machine translation, multilingual NLP, and AI fairness, this study analyzes manifestations across gendered stereotyping, religious oversimplification, regional framing, and cultural normalization; and then synthesizes detection methods, including benchmark-based evaluation, contrastive probing, embedding association tests, and human-in-the-loop assessment. For mitigation, this paper proposes a five-layer framework spanning data auditing, model adaptation, inference-time intervention, post-editing, and governance. To validate the framework, we conduct five proof-of-concept experiments: cross-lingual gender bias detection with statistical testing, systematic cultural fidelity evaluation under prompt engineering, contrastive sentiment analysis under high-/low-risk contexts, word embedding association tests (WEAT) with permutation-based significance, and an integrated audit pipeline with automated mitigation. Results demonstrate significant gender bias (χ²=29.99, p<0.001), a pervasive "male-as-default" phenomenon, significant gains from culture-aware prompting (p=0.03), and robust embedding-space bias (permutation test p=0.0001). The audit pipeline successfully integrates detection and mitigation into an actionable workflow. We conclude by outlining future directions for low-resource languages, intersectional bias, and production-level deployment.