This repository contains the source code for the paper RASP: Robust Mining of Frequent Temporal Sequential Patterns under Temporal Variations.
In this work, we propose RASP, an algorithm for Robust and resource-Adaptive mining of temporal Sequential Patterns. RASP is built upon the following ideas, each devised to address the above limitations:
- Relaxed TSPs and Duplicated Pattern Matching: For robustness against temporal variation, RASP enables multiple TSPs to share the same instance based on the novel concept of a relaxed TSP, which permits a predefined level of time gap deviation.
- Resource-Adaptive Automatic Hyperparameter Tuning: RASP gradually increases the sizes of TSPs to detect larger TSPs. In order to maintain a proper number of TSPs of each size, \method adaptively adjusts thresholds based on the available resources, enhancing its usability.
- Tree-based Concise Data Structure: RASP employs a tree-based compact data structure to efficiently manage the increasing number of TSPs, improving both speed and space efficiency.
All datasets are available at this link.
Experiment | Dataset | Event | Source |
---|---|---|---|
Main | Neuron Activity | Spike of a Neuron | CN2 Simulator |
Additional | E-Commerce | Click on a product | YOOCHOOSE Gmbh |
To install requirements, run the following command on your terminal:
pip install -r requirements.txt
To execute RASP on neuron activity datasets, run this command:
./run.sh
To execute RASP on an e-commerce dataset, run this command:
./run_case.sh
To evaluate the result TSPs, run this command:
python main.py -a read_ndcg_rc_exp
This code is free and open source for only academic/research purposes (non-commercial). If you use this code as part of any published research, please acknowledge the following paper.