Smartness Experiments

This repository contains code to conduct experiments with Regression Models, Non-IID analysis and Federated Learning.

Regression Models:

Experiments with a Wide and Deep neural network to conduct our regression task. Besides this model, this repository contains some notebooks with random forest, xgboost and other algorithms.

Non-IID analysis:

Some analysis was conducted using our datasets to check if our datasets were Non-IID datasets.

Federated Learning:

Federated Learning was developed using Flower framework. The code for Federated Learning is on /smartness-flwr folder. Experiments was conduct using horizontal architecture.

All these experiments was conduct to produce our paper: "An Experimental Framework for Studying Non-IID Data in Federated Learning for Network Telemetry"