Spatiotemporal Event Sequence Mining

Spatiotemporal event sequences represent the sequences of event types whose spatiotemporal instances frequently follow each other in spatiotemporal context. In this work, we present spatiotemporal event sequence mining from spatio-temporal event datasets that contain evolving region trajectories. We propose two algorithms for discovering spatio-temporal event sequences. We formally define a flexible spatio-temporal follow relationship, introduce various data models for capturing the sequence forming behavior. Lastly, we present an extended experimental evaluation that demonstrates the computational efficiency of our algorithms.

See the publication here

Source code

SOLGRIND: Solar Graph Index

A graph-based index for indexing spatiotemporal relationships appearing among the trajectory-based solar event instances. For the source code and dataset check our project repository in BitBucket.

Measuring Significance for Spatiotemporal Co-occurrences

Development of a new significance measure for measuring the significance of spatiotemporal co-occurrences: the J* measure. For the source code and dataset check our project repository in BitBucket.

Indexing Solar Event Data for Mining STCOPs

We have developed a framework for mining spatiotemporal co-occurrence patterns. Two well-known trajectory-based indexing techniques are successfully utilized and the mining process, now, can be done even in real-time for solar data.

The raster data collected from Solar Dynamic Observatory is huge. 17 feature finding teams and their modules generate vector data. Using this vector data, we perform a data-driven analysis of solar phenomena. However, previously used database settings were not satisfactory and feasible for large-scale data analysis. In this project, we have indexed solar vector data for mining solar co-occurrences, in order to increase efficiency.

Indexing techniques, we have utilized for the STCOP-mining framework are the Scalable and Efficient Trajectory Index (SETI) and Chebyshev Polynomial Index (CPI). The framework mimics a simple database environment using binary files. The source code for this project can be found on BitBucket. The framework is implemented in C++, using Boost libraries.

Mining Spatiotemporal Co-occurrence Patterns in Non-relational Databases

In this research project, we aim to create a system that can mine spatiotemporal co-occurrence patterns in distributed non-relational databases.

In the era of big spatiotemporal data, it is unavoidable to use distributed non-relational databases for storing spatiotemporal event instances.

Current relational database management systems offer rich functionality for geometric operations (e.g. PostGIS); but, for mining massive spatiotemporal datasets, newly popular non-relational database systems can perform much better. In this project, we designed data models for storing spatiotemporal event instances and implemented spatiotemporal join procedures for mining spatiotemporal co-occurrence patterns. Our system uses the Accumulo database and a Java client for the implementation of mining algorithms.

Spatiotemporal Frequent Pattern Mining from Solar Datasets

We surveyed the current algorithms on spatiotemporal frequent pattern mining algorithms using the solar event datasets. Moreover, we exhibit possible future directions for knowledge discovery from solar event datasets.

The current work on spatiotemporal frequent pattern mining that can be used in solar event datasets includes spatiotemporal co-occurrence patterns (STCOP) and spatiotemporal event sequences (STES).

We have also presented possible future work that can be used for solar event datasets that are:

  • Periodic Patterns

  • Convergence Patterns

  • Propagation Patterns

See this publication here

Trajectory Modeling and Indexing in NoSQL Databases

With the ever-growing nature of spatiotemporal data, it is inevitable to use non-relational and distributed database systems for storing massive spatiotemporal datasets. In this chapter, the important aspects of non-relational (NoSQL) databases for storing large-scale spatiotemporal trajectory data are investigated. Mainly, two data storage schemata are proposed for storing trajectories, which are called traditional and partitioned data models. Additionally spatiotemporal and non-spatiotemporal indexing structures are designed for efficiently retrieving data under different usage scenarios. The results of our experiments exhibit the advantages of utilizing data models and indexing structures for various query types. Our book chapter can be found here.

ERMO-DG: Evolving Region Moving Object Random Dataset Generator

A random spatiotemporal dataset generator that is intended for generating spatiotemporal co-occurrence patterns. Large-scale spatio-temporal data is crucial for testing the spatiotemporal pattern mining algorithms. ERMO-DG is a highly parameterized dataset generator that can be utilized to create spatiotemporal instances of event types with certain characteristics. The documentation can be found here

Spatiotemporal Frequent Pattern Mining from Evolving Region Trajectories

This SpringerBrief provides an overview within data mining of spatiotemporal frequent pattern mining from evolving regions to the perspective of relationship modeling among the spatiotemporal objects, frequent pattern mining algorithms, and data access methodologies for mining algorithms. While the focus of this book is to provide readers insight into the mining algorithms from evolving regions, the authors also discuss data management for spatiotemporal trajectories, which has become increasingly important with the increasing volume of trajectories.

This brief describes state-of-the-art knowledge discovery techniques to computer science graduate students who are interested in spatiotemporal data mining, as well as researchers/professionals, who deal with advanced spatiotemporal data analysis in their fields. These fields include GIS-experts, meteorologists, epidemiologists, neurologists, and solar physicists.

See the book here