Topical Behavior Prediction from Massive Logs

TL;DR

This paper introduces a deep learning framework combining RNN and CNN architectures to predict topical behavior from large-scale network logs, leveraging spatial and temporal relationships for improved accuracy.

Contribution

It proposes a novel spatially connected convolutional network (SCCN) framework and data reduction techniques for effective topical behavior prediction.

Findings

Temporal and spatial modeling improves prediction accuracy.

SCCN outperforms traditional MLP networks.

Spatial and temporal features contribute to better results.

Abstract

In this paper, we study the topical behavior in a large scale. We use the network logs where each entry contains the entity ID, the timestamp, and the meta data about the activity. Both the temporal and the spatial relationships of the behavior are explored with the deep 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 the CNN, we propose several reduction steps to form the topical metrics and to place them homogeneously like pixels in the images. The experimental result shows both temporal and spatial gains when compared against a multilayer perceptron (MLP) network. A new learning framework called the spatially connected convolutional networks (SCCN) is introduced to predict the topical metrics more efficiently.