Related Experiment Video
Updated: Jan 25, 2026

Mass Spectrometry-Guided Genome Mining as a Tool to Uncover Novel Natural Products
Published on: March 12, 2020
GatewayNet: a form of sequential rule mining
Phillip C S R Kilgore1, Nadejda Korneeva2,3, Thomas C Arnold3
1Department of Computer Science, LSU Shreveport, 1 University Place, Shreveport, 71115, USA. pkilgore@lsus.edu.
The new GatewayNet algorithm analyzes event sequences to test the gateway hypothesis of drug abuse. While it found causal gateway rules in synthetic data, empirical data showed no such rules but did identify transitions into benzodiazepine use.
Area of Science:
- Computer Science
- Data Mining
- Substance Abuse Research
Background:
- The gateway hypothesis, suggesting developmental stages in drug abuse, has been debated since the 1970s.
- Previous research yielded contradictory findings regarding the gateway hypothesis.
- A novel algorithm is proposed to detect association and causation in event sequences.
Purpose of the Study:
- To develop and validate an algorithm for assessing the gateway hypothesis.
- To introduce GatewayNet, a refinement of sequential rule mining for "initiation rule mining" and causal inference.
- To apply GatewayNet to both synthetic and real-world substance use data.
Main Methods:
- Developed GatewayNet, an initiation rule mining algorithm, to infer causal relationships from event sequences.
- Constructed a transaction database using a first-order Markov chain for gateway network generation.
- Evaluated GatewayNet on de-identified urine drug screening data from August 1998 to June 2011.
Main Results:
- GatewayNet identified a chain of causal gateway rules in synthetic data.
- No evidence of gateway rules was found in the empirical substance use data.
- Two documented transitions into benzodiazepine use were identified in the empirical data.
Conclusions:
- GatewayNet shows potential for analyzing substance use data and other event sequence data.
- Future work includes optimizing GatewayNet for improved speed and performance.
- The algorithm offers a new approach to understanding complex event sequences and potential causal links.
More Related Videos
Related Concept Videos
Indeterminate Forms and L’Hôpital’s Rule
Exceptions to the Octet Rule
Lewis Symbols and the Octet Rule
The Aufbau Principle and Hund's Rule
The Quotient Rule
Midpoint Rule

