Analyzing Service Order Data using Association Mining

Association mining was originally developed and successfully applied to analyze the sale transactions of retail goods. According to common practices of association mining, each item is treated equally. But in a service order transaction especially in a telecom setup, an item (service) can be in several states. We propose a simple solution to analyze service order data using association mining by tagging items with corresponding state identifiers. We successfully apply our result to an association mining based product recommender system. Results from random and näıve product recommender systems are also reported for the comparison purposes.