Platform Management System Host-Based Anomaly Detection using TF-IDF and an LSTM Autoencoder
Bibliographic record
Abstract
Supervisory Control and Data Acquisition (SCADA) systems are at the core of many types of critical infrastructure and have become high value targets for cyber attack. SCADA systems are designed to be both available and reliable. With the identification of possible vectors for cyber attack there is a need for monitoring these systems for malicious behaviour. One type of SCADA network is a Platform Management System (PMS) that enables the centralized control and monitoring of numerous subsystems.The aim of this research is to determine the effectiveness of natural language processing and deep learning techniques in detecting host-based anomalies within a PMS network. Effectiveness is determined through the metrics derived from the confusion matrix. System monitor (Sysmon) logs are collected from the PMS subsystem hosts and features are extracted from these host logs using Term Frequency – Inverse Document Frequency (TF-IDF). A Long Short-Term Memory (LSTM) autoencoder is used to detect anomalies.In order to achieve this aim, a pipeline was developed for anomaly detection. Host logs were collected from subsystems on the PMS network and processed using TF-IDF, then used to train the LSTM model. Once trained, new attack data was introduced to the anomaly detection pipeline to determine the effectiveness of NLP and deep learning techniques to detect host-based anomalies within the PMS network.Performance metrics of this anomaly detection pipeline are presented, including accuracy, precision, recall, F1 Score and Matthew’s Correlation Coefficient (MCC). The results of this research show that the proposed pipeline can detect a range of cyber attacks occurring on the PMS network.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".