What question did this study set out to answer?

The aim is to develop a lightweight query routing system for retrieval-augmented dialogue (RAG) systems.

April 17, 20260 citationsOpen Access

Lightweight Query Routing for Retrieval-Augmented Dialogue: A TF–IDF and Logistic Regression Classifier with Confidence-Based Clarification

Key Points

The aim is to develop a lightweight query routing system for retrieval-augmented dialogue (RAG) systems.
Implemented a five-way query router: NO_RAG, LOCAL_RAG, WEB_RAG, MEMORY_ONLY, ASK_CLARIFICATION.
Utilized TF-IDF and multinomial logistic regression for routing.
Set a confidence threshold of 0.45 for fallback.
Conducted 5-fold cross-validation (CV) for performance assessment.
Achieved 93.6% accuracy on test data.
Obtained a macro F1 score of 0.935.
Reached a weighted ROC AUC of 0.993 on balanced examples.

Abstract

A lightweight five-way query router for RAG-based dialogue systems using TF–IDF + multinomial logistic regression. Routes user queries to NORAG, LOCALRAG, WEBRAG, MEMORYONLY, or ASKCLARIFICATION with a 0. 45 confidence threshold fallback. Achieves 93. 6% test accuracy, 0. 935 macro F1, and 0. 993 weighted ROC AUC on 470 balanced examples. Includes full pipeline, Streamlit demo, confusion matrices, and 5-fold CV.

KI fragen

Bookmark

View Full Paper

Cite This Study

Anuj Yadav (Wed,) studied this question.

synapsesocial.com/papers/69e1d0165cdc762e9d8592ce https://doi.org/https://doi.org/10.5281/zenodo.19588993

KI fragen

Bookmark

View Full Paper