Deep Temporal Point Process Approach to Packet Stream Analysis
Document Type
Thesis
Degree Name
Master of Science (MS)
Department
Computer Science and Info Sys
Date of Award
5-22-2024
Abstract
In this work, we perform packet stream analysis using the Deep Temporal Point Process. The Deep Temporal Point process is a probabilistic generative model, which is used as a model that learns a pattern of a series of events and predicts the occurrence of future events (Saha et al., 2019). Instead of using this model as a generative model, we intend to use it for classification through learning event sequence patterns. To perform packet stream analysis, we use the packet arrival time among the raw data of the packets, which is regarded as the event occurrence time from the TPP point of view and learned. The time sequence input data is output as a probability distribution through the intensity function. Data from one attack category is trained for each model, and the classification task is performed by comparing the difference in likelihood from other models. Finally, we compare the performance between the used model and the new model that applies NTPP to packet stream analysis.
Advisor
Jinoh Kim
Subject Categories
Computer Sciences | Physical Sciences and Mathematics
Recommended Citation
KIM, Jinpyo, "Deep Temporal Point Process Approach to Packet Stream Analysis" (2024). Electronic Theses & Dissertations. 1172.
https://digitalcommons.tamuc.edu/etd/1172
