Edge_Intent_Classifier // Milestone 5 (Architecture Spec)

FUTURE FEATURE

Localized Machine Learning Routing Pipeline

The edge classifier architecture converts raw natural language commands into mathematical token blocks, training a lightweight regression model to route inputs locally before querying heavier LLM endpoints.

// Vectorization Engine

Using Scikit-Learn feature vectorizers to map command strings into structured numerical matrices.

// Logical Routing

Logistic Regression classification models to dispatch intents directly to smart home automation scripts.

Status: Architecture specified. Awaiting Phase 3 training corpus collection from local household telemetry logs.