RT-IoT2022 – Feature Guide (Regression Target: flow_duration)

This document is to give insight in the dataset: what columns we’re dealing with and some non-technical explanations for non-technical stakeholders.

1. Summary of the dataset
2. Simple explanation of the dataset
3. Simple metaphor
4. Column index raw data
5. Feature explanation list for non-technical stakeholders
6. Column index cleaned dataset

Summary of the dataset

• Source: https://archive.ics.uci.edu/dataset/942/rt-iot2022
• Licence: Creative Commons Attribution 4.0 International (CC BY 4.0)
• Target: flow_duration
• Rows full file: 123117, (preview .xls file): 500
• Columns: 84

Simple explanation of the dataset

Imagine: each row in the dataset is a communication session between two devices in an IoT network (for example, a smart lamp communicating with a server).

In each session we measure things like:

• How long the session lasted → this is our target (flow_duration).
• How many messages were sent forward and backward → e.g., fwd_pkts_tot and bwd_pkts_tot.