This curated list compiles recent articles on evaluation and optimization models, covering Data Envelopment Analysis, ANP, VIKOR, various grey analysis methods, competition ranking techniques, case studies, linear and integer programming, and a range of optimization algorithms with Python examples.
While many claim that machine learning is merely statistics with a flashy veneer, this article explores the nuanced distinctions between the two fields—examining their goals, methodologies, and examples such as linear regression—to clarify why they are related yet fundamentally different.
This article introduces the Transformer architecture, explains its attention mechanism, details its adaptations for natural language processing, computer vision, and OCR tasks, and presents experimental results of various models such as BERT, ELECTRA, Swin Transformer, and CRNN-BCN on large-scale medical data from Haodf.com.
This article explains the fundamental assumptions of linear regression, compares linear and nonlinear models, discusses multicollinearity, outliers, regularization, heteroscedasticity, VIF, stepwise regression, and reviews essential evaluation metrics such as MAE, MSE, RMSE, R² and Adjusted R².
The article explains classification as a data‑analysis task that builds models to assign new observations to predefined categories, outlines its implementation steps, describes various data types (boolean, nominal, ordinal, continuous, discrete), presents common machine‑learning classifiers such as decision trees and neural networks, and highlights practical applications like crime detection, disease risk prediction, and credit assessment.
Taobao’s rapidly expanding technical data system faced naming inconsistencies, low table reuse, and costly, inefficient data usage, prompting a joint effort with DataWorks to digitize model evaluation, enforce standardized governance, deliver intelligent end‑to‑end modeling tools, and launch a development assistant, resulting in a health‑monitoring dashboard, upgraded data maps, and a roadmap for further automation and architecture refinement.
This presentation explores how to scientifically define, apply, evaluate, and interpret data in financial risk management, covering data alignment with business goals, feature selection, model metrics such as KS and PSI, handling pandemic effects, and methods for model explainability.
This presentation explores how to scientifically define, apply, evaluate, and interpret data in financial risk management, covering data alignment with business goals, feature selection, model metrics like KS and PSI, handling pandemic impacts, and methods for model explanation and improvement.
This article introduces CatBoost, outlines its key advantages such as automatic handling of categorical features, symmetric trees, and feature combination, and provides a step‑by‑step Python tutorial—including data preparation, model training, visualization, and feature importance analysis—using a CTR prediction dataset.
This article introduces the fundamentals of transfer learning, explains its theoretical foundations and formulas, and demonstrates how multi‑task learning and domain‑adaptation techniques are applied to financial risk‑control scenarios to overcome label scarcity, distribution shift, and model complexity challenges, presenting detailed experimental results and analysis.
The Hello algorithm testing team outlines its workflow—from data collection and cleaning through model training, evaluation, and deployment—while teaching machine‑learning fundamentals, detailing company‑wide use cases, defining key terms, and describing four testing capability dimensions covering data quality, service reliability, model performance, and system engineering.
This article explains how an intelligent risk control system can achieve fully automated, end-to-end model iteration, detailing the multi-layer architecture, sample and feature selection, automated training, evaluation, scoring, deployment, and the efficiency gains compared with manual processes.
This article provides a comprehensive summary of machine learning model evaluation metrics, covering accuracy, precision, recall, F1, RMSE, ROC/AUC, KS test, and scoring cards, with explanations, formulas, code examples, and practical considerations for model performance assessment.
The article reviews the most important additions in scikit‑learn 0.24, including faster hyper‑parameter search methods, ICE plots, histogram‑based boosting improvements, new feature‑selection tools, polynomial‑feature approximations, a semi‑supervised classifier, MAPE metric, enhanced OneHotEncoder and OrdinalEncoder handling, and a more flexible RFE interface.
This tutorial walks through a practical case of predicting employee attrition, demonstrating how to assess and compare machine‑learning models using ROC AUC, confusion matrices, precision‑recall trade‑offs, and the Evidently library to generate performance dashboards, helping choose the best model for production.
DiDi’s shift to machine‑learning‑driven ride‑hailing services revealed major QA challenges—data and feature quality, model verification, and API stability—prompting a four‑pillar framework and a unified “Strategy‑Center 1.0” platform to systematically monitor, evaluate, and improve model effectiveness, bias paths, and feature discovery.
This tutorial walks through a complete workflow for assessing machine‑learning models—using a Kaggle HR attrition dataset, comparing Random Forest and Gradient Boosting via ROC‑AUC, precision, recall and segment analysis with the Evidently library—to decide which model is ready for production deployment.
This article presents a comprehensive study on C‑side user identity recognition for internet platforms, addressing cold‑start and sample‑scarcity challenges by comparing keyword matching, XGBoost, TextCNN, a fusion model, and an improved Wide & Deep architecture, showing that the latter achieves the highest F1 score of 80.67%.
The talk explores the current pain points of AI testing, outlines data‑quality analysis methods, highlights critical ETL and model‑testing considerations, and shares practical case studies and platform designs to improve machine‑learning quality assurance.
This article explores the definition of a model, distinguishes business, data, and function models, discusses criteria for a good model—including performance, fidelity to real‑world relationships, and interpretability—and examines why a universal model does not exist, all within the context of data science and AI.
This comprehensive guide examines real‑world text classification projects, covering label taxonomy design, data scarcity solutions, efficient annotation, new‑class discovery, algorithm selection, evaluation metrics, OOV handling, model evolution, rule‑model integration, performance‑boosting tricks, and inference under resource constraints.
This article presents a hand‑drawn, illustrated walkthrough of the entire machine‑learning pipeline—from dataset definition, exploratory data analysis, preprocessing, and data splitting to model building, algorithm selection, hyper‑parameter tuning, feature selection, and evaluation for both classification and regression tasks.
In this article, Wang Zhe addresses fifteen common questions about recommendation systems, covering topics such as building cross‑domain knowledge, the role of deep reinforcement learning, handling sparse or low‑sample data, offline‑online evaluation, knowledge graphs, graph neural networks, model interpretability, large‑scale ID embedding, and career advice for engineers.
By applying XGBoost‑based regression models that are retrained daily on two‑week order data and tuned per sub‑package, the mobile‑game recharge monitoring system predicts 10‑minute order volumes, sharply cuts false alarms from hundreds to dozens, and delivers precise, scalable alerts for game operations.
This article explores how deep‑learning‑based image dewarping techniques, from traditional hardware methods to modern U‑Net, Stacked U‑Net and Dilated U‑Net architectures, can correct warped document photos, improve OCR accuracy, and support intelligent verification in high‑throughput business scenarios.
This article presents a comprehensive approach to testing, automating, and monitoring AI prediction models in a high‑traffic environment, covering background, challenges, evaluation metrics, data sampling methods, automated test scripts, and online monitoring to ensure model accuracy, performance, and reliability.
This article explains how to build a credit scoring card model using logistic regression, detailing the full development workflow—from variable selection and WOE encoding to model training, evaluation metrics like AUC and KS, and deployment considerations for risk management.
This article explains how to construct credit scoring card models using WOE‑encoded variables and logistic regression, describes three card types for different credit scenarios, and details an eleven‑step development workflow including variable selection, modeling, scoring, and performance monitoring.
This article presents a detailed case study of building a purchase‑user prediction model by integrating Convolutional Neural Networks for feature extraction with Long Short‑Term Memory networks for time‑series forecasting, covering background, model structure, data augmentation, experimental results, and business impact.
This article provides a comprehensive guide to using Python's scikit-learn library for supervised learning, focusing on linear regression, covering theoretical background, environment setup, data preprocessing, model training, evaluation with mean squared error, cross‑validation, and detailed code examples.
This article explains that there is no universal solution for selecting machine learning algorithms and outlines practical factors—such as data characteristics, problem type, business constraints, and algorithm complexity—to help practitioners systematically narrow down and pick the most suitable models.
The article demonstrates how a simple fully‑connected MLP can be trained on a small 64×64×3 cat‑vs‑non‑cat dataset, achieving perfect training accuracy but only 78 % test accuracy, and explains that parameter explosion, vanishing gradients, and lack of spatial invariance limit MLPs, motivating the shift to CNNs.
This beginner-friendly guide walks through core machine‑learning concepts—from data organization and feature design to supervised and unsupervised algorithms such as perceptron, logistic regression, decision trees, LDA, and ensemble techniques—while explaining model evaluation, overfitting, and practical tuning strategies.
In this interview, Baidu QATC chair Yang Fei discusses his career, the evolving scope of quality assurance from code to AI model testing, key challenges such as service quality and model interpretability, practical approaches for defect discovery, continuous evaluation pipelines, and advice for QA professionals' personal growth.
This article walks through a complete machine‑learning workflow on the UCI Census Income dataset, covering data exploration, preprocessing (including log‑transformation and scaling), model training with Naïve Bayes, Decision Tree and SVM, performance evaluation, hyper‑parameter tuning via grid search, feature importance analysis, and feature selection, providing code snippets and visualizations.
This article uses realistic Hulu business scenarios to illustrate the pitfalls of relying solely on accuracy, precision, recall, RMSE, and other single metrics, and explains how combining complementary evaluation measures such as average accuracy, precision‑recall curves, ROC, F1‑score, and MAPE can provide a more comprehensive assessment of classification, ranking, and regression models.
From foundational concepts to practical examples, this guide walks engineers through ten essential supervised and unsupervised machine‑learning algorithms—decision trees, Naïve Bayes, linear regression, logistic regression, SVM, ensemble methods, clustering, PCA, SVD, and ICA—explaining their theory, real‑world uses, and why they matter.
This article explains the fundamentals of ROC curve construction, the calculation of AUC, compares ROC with PR curves, and provides step‑by‑step examples—including a medical diagnosis scenario and threshold adjustments—to help readers accurately evaluate binary classification models.
The article traces the history of software engineering as a continual shortening of release cycles, discusses the persistent emergence of new development models—from the waterfall to spiral and iterative approaches—and highlights George Box’s insight that while all models are imperfect, the useful ones illuminate practice.
This guide introduces beginners to core machine learning concepts, covering feature design, supervised and unsupervised methods such as perceptron, logistic regression, decision trees, LDA, and ensemble techniques like bagging and boosting, while explaining model evaluation, overfitting, and practical optimization strategies.
This article provides a comprehensive introduction to data mining, covering its definition, goal setting, data sampling, exploration, preprocessing, pattern discovery, model building, evaluation methods, and the main analytical techniques such as classification, regression, clustering, association rules, feature and deviation analysis, and web mining.
In this talk, Quora’s VP of Engineering Xavier Amatriain shares practical lessons from building the company’s large‑scale recommender system, covering data richness, implicit signals, model choices, feature engineering, evaluation strategies, and why distribution isn’t always required.
This presentation details how Ctrip's hotel R&D team leverages machine learning and big data to address OTA-specific challenges, improve key service KPIs, evaluate project benefits, and deploy models through a robust pipeline and architecture, offering practical case studies and operational insights.
Airbnb’s trust and safety team uses a series of machine‑learning models—starting from defining the prediction target, through data sampling and feature engineering, to evaluating precision and recall—to identify and mitigate fraud risks such as chargebacks across its global peer‑to‑peer rental platform.
This article explains how Airbnb's Trust & Safety team builds machine‑learning models to detect fraudulent behavior, covering problem definition, role‑based sampling, feature design techniques such as normalization and CP‑coding, and the trade‑offs between precision and recall in model evaluation.
