Datasets
The ISOT Lab has collected through different projects various datasets some of which are available for public sharing. The following datasets are available:
ISOT Botnet Dataset
The ISOT Botnet dataset is the combination of several existing publicly available malicious and non-malicious datasets.
Click here to download the (2010) ISOT Botnet dataset. Other link for the same dataset.
The UNB ISCX Intrusion Detection Evaluation DataSet is available at http://www.iscx.ca/dataset
ISOT Mouse Dynamics Dataset
The ISOT mouse dynamics dataset consists of mouse dynamics data for 48 users collected over several months.
The dataset cannot be downloaded directly. Instead you need first to fill an agreement about how the data will be used; the agreement has to be signed by a supervisor. Please send the signed agreement to Dr. Issa Traore <itraore at ece.uvic.ca>.
After approving the request, a link will be sent to you to download the dataset with the instructions about how to use it.
Mouse Dynamic Dataset Agreement Template
ISOT Stylometry Datasets
The ISOT stylometry datasets include the following:
ISOT Twitter dataset
- Dataset description
- The dataset consists of a file named “data.zip” contains 100 files, which contains 100 files, each corresponding to one of the (100) authors
- The file “TweetCrawling.zip” contains Java source code to retrieve a JSON structure for a specific Tweet ID
ISOT Forgery dataset
- Dataset description
- The file “forgeryDataset.zip” contains our dataset in plain text
Data Preprocessing
- The file Canonicizers_.java contains some filters (e.g., stopwords) used in our research
ISOT Web Interactions Dataset (Mouse/Keystroke/Site Actions)
This dataset consists of mouse dynamics, keystroke dynamics, and site actions collected freely and continuously in web sessions from 24 users. The dataset can be used in building and evaluating web-based continuous authentication models.
The dataset cannot be downloaded directly. Instead you need first to fill an agreement about how the data will be used; the agreement has to be signed by a supervisor. Please send the signed agreement to Dr. Issa Traore <itraore at ece.uvic.ca>.
After approving the request, a link will be sent to you to download the dataset with the instructions about to use it.
Web Interacions (Mouse/Keystroke/Site Actions) Dataset Agreement Template
ISOT Mouse Gestures Dataset
This dataset consists of mouse gesture dynamics datasets, whereby 41 participants were asked to produce gesture samples using the mouse, and the dynamics were captured accordingly. The dataset includes both regular samples and samples produced in adversarial situation (i.e. forgery attempts).
The dataset cannot be downloaded directly. Instead you need first to fill an agreement about how the data will be used; the agreement has to be signed by a supervisor. Please send the signed agreement to Dr. Issa Traore <itraore at ece.uvic.ca>.
After approving the request, a link will be sent to you to download the dataset with the instructions about to use it.
Mouse gesture dynamics Dataset Agreement Template
ISOT Fake News Dataset
The ISOT Fake News Dataset is a compilation of several thousands fake news and truthful articles, obtained from different legitimate news sites and sites flagged as unreliable by politifact.com.
Dataset Description
Click here to download the ISOT Fake News Dataset
ISOT HTTP Botnet Dataset
The ISOT HTTP botnet dataset consists of 2 traffic captures: malicious DNS data for 9 different botnets, and benign DNS for 19 different well known software applications.