Mining association rules from dataset containing predetermined decision itemset and rare transactions
Zongyao Sha, Jiangping Chen
Abstract
Zongyao Sha, Jiangping Chen
Abstract
Association rules may exist in many transaction datasets. It is valuable if those rules can be extracted. During the extraction process, efficiency and effectiveness are the main concerns. This paper proposed the concept of association rule discovery from dataset containing predetermined decision itemset (PDI) and rare transaction. PDI is a set of items that users are interested in, while rare transaction is a subset from a transaction set where items in the subset contains association rules with very high confidence and very low support, and that the rest transactions show zero confidence for the same association rules. Such association rules, due to their low support value, can be easily ignored by traditional association rule mining approaches. We analyzed those two scenarios and presented the corresponding mining algorithm, i.e., ARM-PCI-RT through an application in optimizing producing environment as an example. An optimized producing environment is a key decision-making process in bio-chemical industry production. Due to the complex mechanism of bio-chemical production, understanding the favorable production environment is very difficult. On the other hand, a great amount of data has been accumulated through industry production over years. It is possible to find out valuable association rules that may contribute to the improvement of production efficiency and quality through data mining and knowledge discovery.
OpenAlex reports 4 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Association rules may exist in many transaction datasets. It is valuable if those rules can be extracted. During the extraction process, efficiency and effectiveness are the main concerns. This paper proposed the concept of association rule discovery from dataset containing predetermined decision itemset (PDI) and rare transaction. PDI is a set of items that users are interested in, while rare transaction is a subset from a transaction set where items in the subset contains association rules with very high confidence and very low support, and that the rest transactions show zero confidence for the same association rules. Such association rules, due to their low support value, can be easily ignored by traditional association rule mining approaches. We analyzed those two scenarios and presented the corresponding mining algorithm, i.e., ARM-PCI-RT through an application in optimizing producing environment as an example. An optimized producing environment is a key decision-making process in bio-chemical industry production. Due to the complex mechanism of bio-chemical production, understanding the favorable production environment is very difficult. On the other hand, a great amount of data has been accumulated through industry production over years. It is possible to find out valuable association rules that may contribute to the improvement of production efficiency and quality through data mining and knowledge discovery.
Key concepts: Association rule learning, Database transaction, Computer science, Data mining, Transaction data, Production (economics), Process (computing), Knowledge extraction