Mining association rules from imprecise ordinal data

Categorical data can generally be classified into ordinal data and nominal data. Although there have been numerous studies on finding association rules from nominal data, few have tried to do so from ordinal data. Additionally, previous mining algorithms usually assume that the input data is precise and clean, which is unrealistic in practical situations. Real-world data tends to be imprecise due to human errors, instrument errors, recording errors, and so on. Therefore, this paper proposes a new approach to discovering association rules from imprecise ordinal data. Experimental results from the survey data show the feasibility of the proposed mining algorithm. Performance analyses of the algorithms also show that the proposed approach can discover interesting and valuable rules that could never be found using the conventional approach, the Apriori algorithm.