Discovering Classification Rules from Multi Relational Database Having Multiple Class Values?

Today in the real world data are stored in a structured format known as relational database. An increasing number of data mining applications involve the analysis of complex and structured types of data and require the use of expressive pattern languages. So, to classify object in one relation, we have to collect information from another relation. Multi-relational classification is the process of building a classifier based on information stored in multiple relations. Multi relational classification is used to predict behavior and unknown patterns automatically. This paper presents the various approaches of classification for relational database such as Tuple ID Propagation, Selection Graph, Multi-View Learning etc. and proposed work for multi class classification using Tuple ID Propagation in multi relational database.