Attribute oriented induction with star schema

Attribute-Oriented Induction of High-level Emerging Pattern(AOI-HEP) is a combination of Attribute Oriented Induction (AOI) and Emerging Patterns (EP). AOI is a summarisation algorithm that compact a given dataset into small conceptual descriptions, where each attribute has a defined concept hierarchy. This presents patterns are easily readable and understandable.Emerging patterns are patterns discovered between two datasets and between two time periods such that patterns found in the first dataset have either grown (or reduced) in size, totally disappeared or new ones have emerged. AOI-HEP is not influenced by border-based algorithm like in EP mining algorithms. It is desirable therefore that we obtain summarised emerging patterns between two datasets. We propose High-level Emerging Pattern (HEP) algorithm. The main purpose of combining AOI and EP is to use the typical strength of AOI and EP to extract important high-level emerging patterns from data. The AOI characteristic rule algorithm was run twice with two input datasets,to create two rulesets which are then processed with the HEP algorithm. Firstly, the HEP algorithm starts with cartesian product between two rulesets which eliminates rules in rulesets by computing similarity metric (a categorization of attribute comparisons). Secondly, the output rules between two rulesets from the metric similarity are discriminated by computing a growth rate value to find ratio of supports between rules from two rulesets. The categorization of attribute comparisons is based on similarity hierarchy level. The categorisation of attributes was found to be with three options in how they subsume each other. These were Total Subsumption HEP (TSHEP), Subsumption Overlapping HEP (SOHEP) and Total Overlapping HEP (TOHEP) patterns. Meanwhile, from certain similarity hierarchy level and values, we can mine frequent and similar patterns that create discriminant rules. We used four large real datasets from UCI machine learning repository and discovered valuable HEP patterns including strong discriminant rules, frequent and similar patterns. Moreover, the experiments showed that most datasets have SOHEP but not TSHEP and TOHEP and the most rarely found were TOHEP. Since AOI- iii HEP can strongly discriminate high-level data, assuredly AOI-HEP can be implemented to discriminate datasets such as finding bad and good customers for banking loan systems or credit card applicants etc. Moreover, AOI-HEP can be implemented to mine similar patterns, for instance, mining similar customer loan patterns etc.

PARTIAL COPYRIGHT LICENSE I hereby g r a n t t o Simon Fraser U n i v e r s i t y the r i g h t t o lend my thesis, p r o j e c t o r extended essay ( t h e t i t l e o f which i s shown below)

Computational Intelligence, 1991

The development of efficient algorithms for learning from large relational databases is an important task in applicative machine learning. In this paper, we study knowledge discovery in relational databases and develop an attribute-oriented learning method which extracts generalization rules from relational databases. The method adopts the artificial intelligence "learning-from-examples" paradigm and applies in the learning process an attribute-oriented concept tree ascending technique which integrates database operations with the learning process and provides a simple and efficient way of learning from databases. The method learns both characteristic rules and classification rules of a learning concept, where a characteristic rule characterizes the properties shared by all the facts of the class being learned; while a classification rule characterizes the properties that distinguish the class being learned from other classes. The learning result could be a conjunctive rule or a rule with a small number of Jisjuncts. Moreover, learning can be performed with databases containing noisy data and exceptional cases using database statistics. Our analysis of the algorithms shows that attributeoriented induction substantially reduces the computational complexity of the database learning process.

Rough Set and Knowledge Technology, 2010

Our research devoted to develop a new method of generation of decision rules set. This method is compiled using two different mechanism. The first one is based on applying a new constructive induction algorithm to investigated dataset. The belief networks are used in this algorithm The aim is to find the most important descriptive attribute that is calculated from other attributes. The second part of presented method is improvement algorithm that is used in optimization process of gathered rule set. The results of our research contains comparison of classification efficiency using several datasets.

1992

Knowledge discovery in databases, or data mining, is an important issue in the development of data-and knowledge-base systems. An attribute-oriented induction method has been developed for knowledge discovery in databases. The method integrates a machine learning paradigm, especially learning-from-examples techniques, with set-oriented database operations and extracts generalized data from actual data in databases. An attribute-oriented concept tree ascension technique is applied in generalization, which substantially reduces the computational complex@ of database learning processes. Different kinas of knowledge rules, including characteristic rules, discrimination rules, quantitative rules, and data evolution regularities can be discovered efficiently using the attribute-oriented approach. In addition to learning in relational databases, the approach can be applied to knowledge discovery in nested relational and deductive databases. Learning can also be performed with databases containing noisy data and exceptional cases using database statistics. Furthermore, the rules discovered can be used to query database knowledge, answer cooperative queries and facilitate semantic query optimization. Based upon these principles, a prototyped database learning system, DBLEARN, has been constructed for experimentation.

Studies in Computational Intelligence, 2005

In this paper we analyze an attribute-oriented data induction technique for discovery of generalized knowledge from large data repositories. We employ a fuzzy relational database as the medium carrying the original information, where the lack of precise information about an entity can be reflected via multiple attribute values, and the classical equivalence relation is replaced with relation of the fuzzy proximity. Following a well-known approach for exact data generalization in the ordinary databases [1], we propose three ways in which the original methodology can be successfully implemented in the environment of fuzzy databases. During our investigation we point out both the advantages and the disadvantages of the developed tactics when applied to mine knowledge from fuzzy tuples.

Information Reuse and Integration, 2001

Attribute-oriented induction approach (AOA) has been developed for knowledge discovery in large relational database. Several kinds of knowledge, such as characteristic rules and discrimination or classification rules can be discovered. These rules may contain unnecessary conditions and/or unnecessary condition- values. A Tuple-oriented approach (TOA) examines one tuple at a time since there are large number of possible combination in such

International Journal of Information Sciences and Techniques, 2013

Classification is widely used technique in the data mining domain, where scalability and efficiency are the immediate problems in classification algorithms for large databases. We suggest improvements to the existing C4.5 decision tree algorithm. In this paper attribute oriented induction (AOI) and relevance analysis are incorporated with concept hierarchy"s knowledge and HeightBalancePriority algorithm for construction of decision tree along with Multi level mining. The assignment of priorities to attributes is done by evaluating information entropy, at different levels of abstraction for building decision tree using HeightBalancePriority algorithm. Modified DMQL queries are used to understand and explore the shortcomings of the decision trees generated by C4.5 classifier for education dataset and the results are compared with the proposed approach.

Journal of Advanced Computational Intelligence and Intelligent Informatics, 2007

In this paper we focus on the induction of classification rules from examples. Conventional algorithms fail in discovering effective knowledge when the database contains irrelevant information. We present a new rule extraction method, RGT, which tackles this problem by employing only relevant and irredundant attributes. Simplicity of rules is also our major concern. In order to create simple rules, we estimate the purity of patterns and propose a rule enlargement approach, which consists of rule merging and rule expanding procedures. In this paper, we describe the methodology for the RGT algorithm, discuss its properties, and compare it with conventional methods.

2009

Supervised classification is a spot/task of data mining which consists in building a classifier from a set of examples labeled by their class (learning step) and then predicting the class of new examples with a classifier (classification step). In supervised classification, several approaches were proposed such as: Induction of Decision Trees, and Formal Concept Analysis. The learning of formal concepts is based, generally, on the mathematical structure of Galois lattice (or concept lattice). The complexity of generation of Galois lattice, limits the application fields of these systems. In this paper, we present several methods of supervised classification based on Formal Concept Analysis. We present methods based on concept lattice or sub lattice. We also present the boosting of classifiers, an emerging technique of classification. Finally, we propose the boosting of formal concepts: a new adaptive approach to build only a part of the lattice including the best concepts. These concepts are used as classification rules. Experimental results are given to prove the interest of the proposed method.