Summarized Network Behavior Prediction

TL;DR

This paper introduces a novel framework combining RNN and CNN architectures to analyze and predict entity-wise network behavior by leveraging both temporal and spatial relationships in network logs.

Contribution

It presents a new learning framework called spatially connected convolutional networks (SCCN) that effectively captures spatial and temporal behavior patterns in network data.

Findings

SCCN outperforms traditional MLP in behavior prediction.

Spatial and temporal features both improve prediction accuracy.

The approach effectively transforms behavioral data into image-like formats for CNN processing.

Abstract

This work studies the entity-wise topical behavior from massive network logs. Both the temporal and the spatial relationships of the behavior are explored with the learning architectures combing the recurrent neural network (RNN) and the convolutional neural network (CNN). To make the behavioral data appropriate for the spatial learning in CNN, several reduction steps are taken to form the topical metrics and place them homogeneously like pixels in the images. The experimental result shows both the temporal- and the spatial- gains when compared to a multilayer perceptron (MLP) network. A new learning framework called spatially connected convolutional networks (SCCN) is introduced to more efficiently predict the behavior.