A Bag-of-Features Framework for Time Series Classification

This is a supporting page to our paper -  A Bag-of-Features Framework to Classify Time Series

by Mustafa Gokce Baydogan, George Runger and Eugene Tuv 
*this study is presented in INFORMS 2011@Charlotte. The presentation is available here."
*the technical report related to this study is available in Technical Reports category in Files section."
"the paper submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI) will be available in Papers category in Files section."
 
CODES
The codes are provided in the files section. Here is the direct link to the folder. You will find zip file containing:
  1. R implementation: We use the "Random Forest" package in R so you need to install R software (http://www.r-project.org/) and then install the required library using the command install.packages("randomForest").  For parallel implementation of TSBF for multicore computers, install "doMC" package and modify the number of cores to be used accordingly.
  2. C code and compiled libraries: The subsequence and codebook generation is implemented in C and called directly from R. The code is compiled using an 64bit Ubuntu 12.04 system (*.so file is in the archive). For Windows (32bit or 64bit) or Linux system (32bit), you need to compile the C code yourself by running "R CMD SHLIB yourpath/extract_sub.c" on your command window (in Windows) or terminal (in Linux).
  3. Wrapper for functions implemented in C: In order to use the functions implemented in C, a wrapper is coded in R. This file is included by the main R implementation. If you are on Windows operating system, R commands to include the compiled library must be changed (must switch to *.dll from *.so after compiling the library in Windows).
  4. Example dataset: GunPoint dataset from UCR time series database is provided as an example.
DATASETS
We tested TSBF on 45 datasets from UCR time series database. Our detailed results are provided in the Results category under Files section. Also summarized version of this information is  available here.

HOW TO RUN TSBF

Please include datasets and codes in the same folder.  The algorithm is run in R software (use R code to run the algorithm).  You may need to set your working directory as the same directory where you store your files if you do not run R in the same directory. If you have any problems running the code, please contact. A detailed description is available here.

Copyright © 2014 mustafa gokce baydogan

LinkedIn
Twitter
last.fm