Tagged articles
XGBoost
51 articles · Page 1 of 1
This article details the end-to-end design of a production-grade AI financial risk control system, covering five-layer architecture, real-time and offline feature engineering with Flink and Spark, a four-layer model stack combining rules, XGBoost, graph neural networks, and LLMs, real-time decision engines using Triton, incremental graph analytics with Nebula Graph, and continuous monitoring with champion-challenger A/B testing.
AI risk controlGraph Neural NetworksLLM-assisted decision
0 likes · 15 min read

Milvus 3.0's search iterator pagination fails after reranking because the cursor uses the last page's score as a boundary, but reranking replaces ANN distances with model predictions (e.g., XGBoost raw scores), causing the filter condition to exclude nothing and return duplicate pages. Milvus now explicitly rejects combining function chains with search iterators rather than fixing pagination, as the fix would require carrying original distances through the entire reduce pipeline—a feature-level change. Workarounds include using offset within single-search limits or moving reranking to the application layer.
MilvusPaginationReranking
0 likes · 21 min read

The article dissects Milvus 3.0's Function Chain, explaining how fine‑grained ranking is moved into the engine with a two‑stage L0 early‑prune and L2 post‑reduction design, native XGBoost scoring, typed protobuf pipelines, and the trade‑offs revealed by source‑code analysis and community discussions.
L0L2Milvus
0 likes · 19 min read

The article announces the winning teams of the 2026 China University Computer Competition Big Data Challenge and provides detailed, step‑by‑step experience reports covering data processing, feature engineering, model design, training strategies, and post‑processing for a cross‑sectional stock ranking task.
LightGBMQuantitative FinanceTime Series
0 likes · 10 min read

This article details a 30‑day AI‑driven shrimp‑farming project built on Alibaba Cloud's Bailei platform, describing data sources, system architecture, model development, daily performance metrics, cost savings, and validation results that demonstrate how AI can replace expert judgment in aquaculture.
AIDockerOpenClaw
0 likes · 16 min read

The article explains how test prediction analytics can replace intuition with data‑driven risk signals, detailing high‑ROI use cases, data governance practices, model selection (favoring XGBoost), and a three‑layer deployment architecture that integrates predictions into CI/CD workflows, backed by concrete results from finance and e‑commerce projects.
CI/CDData‑Driven TestingKubernetes
0 likes · 8 min read

Timber is a multi‑pass compiler that transforms classic ML models such as XGBoost and LightGBM into zero‑dependency C99 binaries, offering microsecond‑level inference latency, HTTP‑compatible serving, and substantial performance gains over Python runtimes, making it ideal for high‑throughput, low‑latency production scenarios.
LightGBMML compilerTimber
0 likes · 8 min read

The article explains how combining test‑prediction analytics with cost‑benefit analysis can turn vague testing intuition into data‑driven decisions, using real‑world case studies, a TCBA decision framework, and concrete guidelines to maximize ROI while reducing test effort.
CI/CDROIXGBoost
0 likes · 8 min read

FinSentLLM integrates multiple LLM experts with structured financial semantic signals, achieving 3‑6% higher accuracy and F1 on the Financial PhraseBank compared to baselines, while DCC‑GARCH and Johansen cointegration analyses confirm a statistically significant long‑term co‑movement between the predicted sentiment signals and stock market dynamics.
DCC-GARCHFinSentLLMJohansen Cointegration
0 likes · 12 min read

This report details a university big‑data project that built a full pipeline—from raw petrochemical market data and text mining to variable selection, XGBoost/LightGBM/Lasoo regression and RNN/LSTM/GRU models—to predict product prices across multiple horizons, evaluate errors, and deliver an interactive demo.
LSTMRNNXGBoost
0 likes · 8 min read

Competing in the 2025 China University Big Data Challenge, our interdisciplinary team leveraged GRU-based time‑series modeling, XGBoost feature fusion, risk‑based clustering, and a custom asymmetric loss to achieve a fourth‑place finish while highlighting the limitations of relying solely on SHSE 300 stock data.
GRUXGBoostensemble learning
0 likes · 5 min read

This article provides a detailed, visual guide to XGBoost's most important hyper‑parameters—such as max_depth, min_child_weight, learning_rate, gamma, subsample, colsample_bytree, scale_pos_weight, alpha, and lambda—explaining how each influences tree complexity, regularization, and model generalization, and offering practical examples for effective tuning.
RegularizationXGBoostgradient boosting
0 likes · 12 min read

This article systematically explores the principles and mathematics behind ARIMA, XGBoost, LSTM, and Transformer models, compares their strengths and weaknesses, and demonstrates a complete end‑to‑end workflow for detecting CPU resource anomalies in a cloud service environment.
ARIMALSTMTransformer
0 likes · 12 min read

Based on a public‑cloud client’s real‑world project, this article details how combining AI large‑model prompting with machine‑learning techniques—first pure large‑model forecasts, then local weighted linear regression, and finally XGBoost—enables automated, accurate investment budget prediction and allocation, reducing analyst workload and scaling to millions of daily calls.
AIInvestment ForecastingLocally Weighted Regression
0 likes · 14 min read

The article details how Bilibili leverages AI algorithms—including XGBoost, statistical rules, XDeepFM, and fine‑tuned SBERT—to optimize VOD transcoding decisions, estimate compute resources and processing time, and analyze live comments, thereby boosting streaming efficiency, utilization, and user experience.
AITranscoding OptimizationXGBoost
0 likes · 19 min read

This article introduces Treelite, an open‑source library that compiles XGBoost, LightGBM, and scikit‑learn tree models into optimized shared libraries, explains its branch‑prediction and comparison‑simplification techniques, and provides step‑by‑step Python examples showing significant inference speed gains across different batch sizes.
LightGBMPythonTreelite
0 likes · 6 min read

This article introduces five essential Python libraries—MLflow, Streamlit, FastAPI, XGBoost, and ELI5—that every junior or intermediate machine‑learning engineer and data scientist should master to streamline experiment tracking, build interactive web apps, deploy models efficiently, achieve fast accurate predictions, and improve model interpretability.
ELI5FastAPIMLflow
0 likes · 8 min read


Didi TechJan 25, 2024 · Artificial Intelligence
Didi’s new Ray‑native XGBoost training platform replaces the fault‑prone Spark solution with a fully Pythonic, fault‑tolerant architecture that leverages Ray’s autoscaling and gang‑scheduling, delivering 2–6× speedups, reduced failure rates, efficient sparse‑vector handling, scalable hyper‑parameter search, and improved resource utilization for large‑scale machine‑learning workloads.
MLOpsRayXGBoost
0 likes · 20 min read

This article outlines the essential concepts, data preparation, model creation, training, deployment, and verification steps for testing machine‑learning models, highlighting dataset requirements, algorithm categories, framework choices, resource considerations, and provides a sample inference request.
AIXGBoostdata preparation
0 likes · 7 min read

This article details a joint research effort between Meituan and East China Normal University that combines cost‑based methods with AI‑driven, data‑centric models to automatically generate and evaluate missing indexes for billions of daily slow queries, improving recommendation accuracy and query performance.
AIIndex RecommendationMySQL
0 likes · 16 min read

The paper proposes a containerized, machine‑learning framework that fuses rule‑based and XGBoost‑driven anomaly detection to monitor daily active users on a cloud music platform, achieving 89 % recall, 81 % precision and up to 74 % recall improvement over traditional threshold methods, while outlining future model refinement and broader metric applicability.
3-sigmaBusiness MetricsHolt-Winters
0 likes · 11 min read

This article explains why post‑hoc interpretation methods such as PDP, ALE, LIME, and SHAP are essential for extracting insights from complex machine‑learning models, demonstrates their mathematical foundations, discusses limitations, and provides a complete Python example using XGBoost on a housing‑price dataset.
ICELIMESHAP
0 likes · 14 min read

This article explains the core concepts of XGBoost as a boosting tree algorithm, describes how it builds ensembles of decision trees to predict outcomes, and provides complete Python implementations for classification and regression using the Scikit-learn interface, along with visualizations of trees and feature importance.
ClassificationPythonXGBoost
0 likes · 4 min read

This article presents a comprehensive overview of the vertical federated XGB algorithm, covering its problem definition, secure multi‑party training techniques, performance‑optimizing oblivious permutation methods, and prediction workflow, while evaluating its scalability and efficiency under various network conditions.
XGBoostprivacy-preserving
0 likes · 12 min read

NetEase Yanxuan built an end‑to‑end time‑series forecasting system for e‑commerce that integrates rich user, product, business and external features with a suite of statistical, machine‑learning and deep‑learning models, delivers predictions via a Tornado‑based service for thousands of SKUs, warehouses, advertising and app traffic, and shows that simpler models like XGBoost often outperform complex deep nets while interpretability and external shocks remain key challenges.
Sales PredictionXGBoostalgorithm system
0 likes · 10 min read


Code DAODec 17, 2021 · Artificial Intelligence
The article explains why XGBoost training can be slow despite its speed focus and presents three acceleration techniques—choosing an optimal tree_method, leveraging cloud resources for larger memory, and using Ray for distributed training—complete with code examples and benchmark results.
RayXGBoostcloud computing
0 likes · 5 min read

Using a road freight accident prediction example, the article warns that interpreting predictive model explanations as causal effects can be misleading, explains when such models may answer causal questions, demonstrates SHAP analysis on an XGBoost model, and recommends causal inference tools like ecoml for reliable effect estimation.
Causal InferenceSHAPXGBoost
0 likes · 10 min read

This article summarizes the author's participation in the Kaggle Jane Street Market Prediction competition, detailing the anonymous feature dataset, utility‑score metric, data preprocessing, the combined AE‑MLP and XGBoost modeling approach, threshold tuning, experimental findings, and references for further study.
KaggleMLPUtility Score
0 likes · 8 min read


58 TechDec 25, 2020 · Artificial Intelligence
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%.
Model EvaluationTextCNNUser Identity
0 likes · 13 min read

This article describes a complete pipeline for recommending automated test scripts for APK releases, covering CSV data preprocessing, feature encoding, tokenization with pkuseg and jieba, and training various machine‑learning models such as LDA, word2vec, XGBoost, deep neural networks, and multi‑label classifiers to predict script execution order.
APK testingModel TrainingXGBoost
0 likes · 14 min read

This article details how Ctrip's search team leverages XGBoost for learning-to-rank, covering L2R concepts, feature engineering, data preparation, model training, hyper‑parameter tuning, offline evaluation, and deployment insights for large‑scale search ranking systems.
XGBoostfeature engineeringlearning to rank
0 likes · 12 min read


58 TechNov 11, 2019 · Artificial Intelligence
The article describes the end‑to‑end architecture, data collection, preprocessing, feature engineering, model selection, training, and hyper‑parameter tuning of 58’s car price estimation platform, which leverages Spark, XGBoost, LightGBM and custom business rules to predict vehicle resale values.
LightGBMXGBoostcar price estimation
0 likes · 11 min read

This article explains how MaFengWo’s recommendation system combines recall, ranking, and rerank stages, details the evolution of its sorting algorithm platform, and shows how data verification and model‑explainability techniques like SHAP and LIME improve online performance and accelerate model iteration.
Model ExplainabilityRecommendation SystemXGBoost
0 likes · 13 min read


Didi TechOct 8, 2019 · Artificial Intelligence
Partnering with Ant Financial, Didi enhanced the open-source SQLFlow platform—translating SQL into end-to-end AI workflows with added deep-learning, XGBoost, clustering and SHAP explanation capabilities and Hive support—to create a “SQL garden” marketplace where analysts can deploy ready-made AI models via simple SQL, speeding enterprise AI adoption.
AISHAPSQLFlow
0 likes · 9 min read


AntTechSep 27, 2019 · Artificial Intelligence
The article describes how Didi's data science team partnered with Ant Financial to co‑build the open‑source SQLFlow platform, enabling analysts to launch AI models via simple SQL, detailing the models contributed, technical extensions, and the broader vision for a universal AI ecosystem.
AISHAPSQLFlow
0 likes · 8 min read

The discrepancy between XGBoost’s Java engine and Spark arose because XGBoost4j treats zero as the default missing value while Spark’s sparse vectors use NaN, causing inconsistent predictions, and was resolved by explicitly setting Float.NaN as the missing value or converting sparse vectors to dense so both engines handle zeros uniformly.
Data EngineeringSparkSparseVector
0 likes · 13 min read

This article outlines the end‑to‑end design of automotive finance risk‑control processes, discusses key data integrity and customer segmentation considerations, and details machine‑learning modeling practices—including logistic regression, decision trees, GBDT, XGBoost, LightGBM and CatBoost—along with an automated platform to streamline model development and deployment.
Credit ScoringGBDTXGBoost
0 likes · 17 min read

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.
Mobile GamingModel EvaluationXGBoost
0 likes · 8 min read

This article shares the experience of Xu An, runner‑up in the 2019 Tencent Advertising Algorithm Competition, detailing practical advice on feature engineering, model selection, efficiency tricks, personal habits, contest rhythm, and learning resources for aspiring participants.
Algorithm ContestLightGBMTencent Advertising Competition
0 likes · 6 min read


58 TechSep 7, 2018 · Artificial Intelligence
The article details how 58.com’s Cupid push control system leverages machine‑learning models, especially XGBoost‑based CTR prediction, to prioritize and filter billions of daily push notifications, improving click‑through rates, reducing user annoyance, and providing a scalable, data‑driven architecture for diverse business services.
AB TestingCTR predictionXGBoost
0 likes · 13 min read

This article describes Ctrip's call‑center staffing challenge, detailing data cleaning, trend analysis, feature engineering, the initial ARIMAX‑Fourier model (V1.0), its limitations, and the improved V2.0 solution that combines TBATS, ARIMA residuals and XGBoost, achieving up to 89.5% prediction accuracy.
Time SeriesXGBoostcall center
0 likes · 9 min read

The article explains how a data analyst can use statistical and machine‑learning models such as linear regression, tree‑based boosting, STL decomposition, and Prophet for both non‑time‑series forecasting and time‑series monitoring, highlighting data‑quality concerns, feature‑engineering practices, and deployment considerations like PMML packaging.
BIProphetSTL
0 likes · 13 min read

This article details a data‑science competition solution that predicts enterprise exit risk using 88 engineered features, LightGBM and XGBoost models with 5‑fold bagging, achieving a top‑25 ranking on the private leaderboard.
Data CompetitionLightGBMXGBoost
0 likes · 5 min read


21CTOSep 20, 2017 · Big Data
This article details a data-driven solution for the 2016 O2O coupon redemption competition, describing dataset partitioning, extensive feature engineering across user, merchant, and coupon dimensions, handling leakage, and model fusion using XGBoost, GBDT, and RandomForest, achieving top AUC scores through weighted ensemble.
GBDTXGBoostcoupon redemption
0 likes · 12 min read

In this interview, a solo female participant from Shanghai Jiao Tong University shares her experience, challenges, and technical insights—including feature engineering, memory management, and model tuning with XGBoost and LightGBM—gained while competing in the Tencent Social Ads algorithm contest.
Model tuningTencentXGBoost
0 likes · 5 min read

The SkullGreymon team, winners of the biggest improvement award in the Tencent Social Ads university algorithm competition, share their journey from a late start in the preliminaries to significant performance gains in the finals, detailing memory‑saving feature extraction techniques, pandas and numpy usage, and their XGBoost modeling approach.
Algorithm CompetitionPandasXGBoost
0 likes · 6 min read

The week‑3 champion team from Peking University shares their experience in the Tencent social advertising algorithm contest, detailing data preparation, feature engineering, and model training strategies that helped them secure the top spot.
Algorithm CompetitionTencentXGBoost
0 likes · 5 min read

The article introduces the Tencent Social Ads university algorithm contest weekly champion team Groot, details their background, and outlines their practical machine‑learning approach—including training set construction, XGBoost model selection, and feature engineering—while encouraging broader participation in such competitions.
AIAdvertisingCompetition
0 likes · 4 min read

The article compiles ten participants' reflections on the Tencent Social Ads university algorithm contest, detailing their data analysis, feature engineering, model selection, challenges faced, and practical advice, while encouraging readers to vote for the most helpful contributions.
AIAlgorithm CompetitionXGBoost
0 likes · 13 min read

The expert live Q&A session for Tencent's Social Ads algorithm competition was recorded and the full FAQ, covering data labeling, feature handling, model selection, imbalance mitigation, and practical engineering tips, is presented here for participants and researchers alike.
AdvertisingCTR predictionCompetition
0 likes · 14 min read

This article explains how to predict customer churn by analyzing competition data, extracting and engineering features, applying GBDT‑based XGBoost models, tuning hyper‑parameters, and improving performance through bagging and stacking model fusion techniques.
Customer ChurnXGBoostmodel fusion
0 likes · 15 min read
