Improving the efficiency of graph-based concept discovery systems

Abay, Nazmiye Ceren
Concept discovery is a process for finding hidden relations from the given set of experiences named as background knowledge [27]. Concept discovery problems are investigated under Inductive Logic Programming (ILP)-based approaches and graph-based approaches [28]. Although ILP-based systems dominate the area, these systems have some problems such as local maxima and local plateaus [15]. Recently, graph based system becomes more popular due to its flexible structure, clear representation of data and ability of overcoming problems of ILP-based systems. Graph based approaches can be classified into two parts defined as structure-based approaches and path-finding approaches according to their methods they use for discovering concepts. The proposed approach can be classified as a combination of both path-finding approaches and methods of association rule mining. It finds paths between the arguments of target instances for concept discovery from the given graph. In addition to path-finding from given graph, association rule mining techniques are used for pruning based on support and confidence. The proposed method is different from the other path-finding approaches because association rule mining techniques are used for reducing search space and finding frequent and strong concept definitions.


Improving scalability and efficiency of ILP-based and graph-based concept discovery systems
Mutlu, Alev; Karagöz, Pınar; Department of Computer Engineering (2013)
Concept discovery is the problem of finding definitions of target relation in terms or other relation given as a background knowledge. Inductive Logic Programming (ILP)-based and graph-based approaches are two competitors in concept discovery problem. Although ILP-based systems have long dominated the area, graph-based systems have recently gained popularity as they overcome certain shortcomings of ILP-based systems. While having applications in numerous domains, ILP-based concept discovery systems still su...
Policy-based memoization for ILP-based concept discovery systems
Mutlu, Alev; Karagöz, Pınar (2016-02-01)
Inductive Programming Logic (ILP)-based concept discovery systems aim to find patterns that describe a target relation in terms of other relations provided as background knowledge. Such systems usually work within first order logic framework, build large search spaces, and have long running times. Memoization has widely been incorporated in concept discovery systems to improve their running times. One of the problems that memoization brings to such systems is the memory overhead which may be a bottleneck. I...
Efficient computation of strong partial transitive-closures
Toroslu, İsmail Hakkı (null; 1993-01-01)
The development of efficient algorithms to process the different forms of the transitive-closure (TC) queries within the context of large database systems has recently attracted a large volume of research efforts. In this paper, we present a new algorithm suitable for processing one of these forms, the so called strong partially-instantiated, in which one of the query's argument is instantiated to a set of constants and the processing of which yields a set of tuples that draw their values form both of the q...
Bergson’s method of intuition: towards a philosophy of life /
Koçkan, Zöhre; Çırakman, Elif; Department of Philosophy (2014)
The purpose of this study is to show how a possible philosophy of life can arise by following Bergson’s method of intuition and to make emphasis on how Bergson’s two fundamental notions (intuition and duration) are capable of grasping the flux of life. The scientific methods, static concepts and classical philosophy are not able to understand the flow of life. Throughout this study it is pointed out a possible philosophy that is able to grasp the flow and the evolution of life. For this aim, Bergson’s metho...
A Counting-Based Heuristic for ILP-Based Concept Discovery Systems
Mutlu, Alev; Karagöz, Pınar; Kavurucu, Yusuf (2013-09-13)
Concept discovery systems are concerned with learning definitions of a specific relation in terms of other relations provided as background knowledge. Although such systems have a history of more than 20 years and successful applications in various domains, they are still vulnerable to scalability and efficiency issues - mainly due to large search spaces they build. In this study we propose a heuristic to select a target instance that will lead to smaller search space without sacrificing the accuracy. The p...
Citation Formats
N. C. Abay, “Improving the efficiency of graph-based concept discovery systems,” M.S. - Master of Science, Middle East Technical University, 2014.