The Mass Estimation Package

Description: This package includes the MATLAB code for Mass Estimation.

References: Kai Ming Ting, Guang-Tong Zhou, Fei Tony Liu and Swee Chuan Tan, Mass Estimation and Its Applications. In: Proceedings of the 16th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 989-998, Washington, DC, 2010. [pdf]

Requirement: MATLAB (>=7.4 R2007a), libsvm.

Attention 1: This package is free for academic usage. You can run it at your own risk. For other purposes, please contact Dr. Kai Ming Ting.

Attention 2: This package was developed by Guang-Tong Zhou. Please feel free to contact him for any problem concerning the code.

This package is also available at https://sourceforge.net/projects/mass-estimation/.

A Step-by-Step Instruction

1. Installation
   • Download the [mass estimation package] (261Kb) and unzip it to a separate directory.
   • Run 'Code/Run_Example_all.m' to obtain simulation results.
   • Build libsvm with 'Library/libsvm/make.m' (for regression tasks only).
   (Notice: we used windows plateform in our experiments. To instal libsvm on other plateforms, please see this page.)


2. Content-based image retrieval
   • Obtain the retrieval data sets and preprocess the data following this instruction [pdf].
   • Place the MATLAB data files in directory 'Data/Retrieval/'.
   • Run 'Code/Run_Retrieval_all.m' to receive the retrieval results, which will be saved in directory 'Results/Retrieval/'.


3. Regression
   • Obtain the regression data sets and preprocess the data following this instruction [pdf].
   • Place the MATLAB data files in directory 'Data/Regression/'.
   • Run 'Code/Run_Regression_all.m' to get the regression results, which will be saved in directory 'Results/Regression/'.


4. Anomaly detection
   • Obtain the anomaly detection data sets and preprocess the data following this instruction [pdf].
   • Place the MATLAB data files in directory 'Data/Anomaly/'.
   • Run 'Code/Run_Anomaly_all.m' to see the anomaly detection results, which will be saved in directory 'Results/Anomaly/'.

More details are described in the readme file. Please excuse us for not providing the data sets due to the copyright issues.