Volume III-2
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., III-2, 53-60, 2016
https://doi.org/10.5194/isprs-annals-III-2-53-2016
© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.
ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., III-2, 53-60, 2016
https://doi.org/10.5194/isprs-annals-III-2-53-2016
© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.

  02 Jun 2016

02 Jun 2016

A RELIABILITY EVALUATION SYSTEM OF ASSOCIATION RULES

Jiangping Chen1, Wanshu Feng1, and Minghai Luo2 Jiangping Chen et al.
  • 1School of Remote Sensing and Information Engineering ,Wuhan University, Wuhan,Hubei,430079,China
  • 2Wuhan Geomatics Institute,Wuhan,Hubei,430022, China

Keywords: Association rules, Evaluation, Reliability, Accuracy, Completeness, Consistency

Abstract. In mining association rules, the evaluation of the rules is a highly important work because it directly affects the usability and applicability of the output results of mining. In this paper, the concept of reliability was imported into the association rule evaluation. The reliability of association rules was defined as the accordance degree that reflects the rules of the mining data set. Such degree contains three levels of measurement, namely, accuracy, completeness, and consistency of rules. To show its effectiveness, the "accuracy-completeness-consistency" reliability evaluation system was applied to two extremely different data sets, namely, a basket simulation data set and a multi-source lightning data fusion. Results show that the reliability evaluation system works well in both simulation data set and the actual problem. The three-dimensional reliability evaluation can effectively detect the useless rules to be screened out and add the missing rules thereby improving the reliability of mining results. Furthermore, the proposed reliability evaluation system is applicable to many research fields; using the system in the analysis can facilitate obtainment of more accurate, complete, and consistent association rules.