Skip to Main Content
Mining is a process of searching data in huge database to infer useful information and deduces relationships and patterns. Though we can predict certain patterns from our database manually, but as soon as size of data increases (becomes in terabytes) it becomes difficult and tedious to deduce the important information from huge database (or data warehouse). Various data-mining algorithms exist to identify possible patterns in data. 1Rule algorithm is one such algorithm but is capable of classifying rule based on only one attribute. This paper extends 1Rule algorithm by classifying more rules based on multiple attributes, so that more accurate decision or prediction are made thereby improving revenue and reducing costs in a company. 1Rule creates a rule for a data based on one attribute, it chooses the rule that gives the lowest classification error after comparing the error rates from all the attributes. In our work we identified and create more rules based on multiple attributes by considering the lowest error attribute as the new classified attribute.