An effective and empirical technique for classification using association rule

Abstract

Organizational policy so called association rule mining and classification are two important data mining terminology and technology in the knowledge discovery process. The combination of these two methods is an important research topic and has many applications in data mining. The combination of these two methods creates a new method called group mining policy or group classification system. The combination of these two methods provides better classification accuracy when classifying data. The research field of content-based data collection requires high efficiency and productivity. Join the mining rule to find the engagement pattern from the data in these applications; we will classify the targets based on the engagement pattern. Our paper focuses on the combination of classification and association rule mining to achieve classified data. In this paper, we propose the use of two new algorithms, CPAR (classification based on predictive association rules) and CMAR (classification based on multi-category association rules), which provide integration and distribution benefits when necessary. CPAR uses anger detection techniques to generate rules directly from training data rather than creating common constructs such as classification entities. Additionally, CPAR develops and tests regulations rather than policy-based procedures to avoid significant regulation. To avoid over-fitting, CPAR evaluates each rule using expected accuracy and uses the top-k rules for prediction. CMAR uses the CR tree model to efficiently store and retain rules in the mined organization and truncate rules based on trust, necessity, and evidence. Distributions were based on a weighted χ analysis using multiple association rules. Extensive experiments show that CMAR is consistent, efficient, and has better mean variance than FOIL (first-person inductive learner) and PRM (predictive rule mining) for classifying different products. The proposed algorithm is better in terms of required memory, time consumption and eliminating intermediate data structure when used.