A complete credit decision system, from scorecard to verified explanation, compiled inside one 0.8B language model (Qwen3.5-0.8B). CC BY-NC 4.0.
-
Updated
Oct 31, 2026 - Python
A complete credit decision system, from scorecard to verified explanation, compiled inside one 0.8B language model (Qwen3.5-0.8B). CC BY-NC 4.0.
ML model for an automatic support tickets prioritization
📉 Plataforma Enterprise de Inteligência de Retenção e MLOps: Streaming (Kafka/Flink), Feature Store (Feast), CatBoost, TreeSHAP, GenAI Copilot com AI Safety e Cockpit em React 19 no Google Cloud Run.
BhoomiSetu is an AI-driven decision intelligence platform built for government agencies and PSUs to transform land acquisition from a reactive bottleneck into a transparent, predictable, and citizen-centric governance workflow for critical infrastructure development.
Logistic z50 ecotone retrieval, iso-NDVI velocity fields and TreeSHAP-attributed GBDT on harmonised Landsat/Sentinel-2 STAC-COG cubes
Geospatial ML platform estimating fair asking prices for Jaipur residential properties, with calibrated conformal intervals and exact multiplicative TreeSHAP factor explanations.
Exact Shapley values, KernelSHAP, path-dependent TreeSHAP and a permutation estimator from first principles, with beeswarm, waterfall and dependence plots
Fast SHAP values: a drop-in Rust engine for the shap library. Same values, 10-80x faster TreeSHAP for XGBoost, LightGBM, random forests and scikit-learn. One line: shap_rs.patch()
Anistroph - multi-domain predictive analytics architecture for temporal and non-temporal modeling, explainability, multidimensional evaluation, and MCP-based inference with Claude and AI agents.
Aplikasi deteksi phishing website menggunakan Random Forest dan TreeSHAP untuk klasifikasi URL serta interpretasi hasil prediksi, dikembangkan dengan Python, Flask, dan Playwright.
SecureSense AI is a Multi-Modal Explainable Trust Intelligence Platform that investigates, verifies, and explains the authenticity of financial communications across emails, SMS, website URLs, PDFs, QR codes, images, and voice recordings using evidence-driven AI.
MSc dissertation (Bristol × UWE): interpretable anomaly detection on the imbalanced UCI-SECOM semiconductor dataset — VAE-GAN augmentation, CatBoost classification, TreeSHAP explanations and a FastAPI real-time demo.
基于 XGBoost + SHAP 的上市公司财务风险预警平台/系统
To associate your repository with the treeshap topic, visit your repo's landing page and select "manage topics."