A Brief Overview of Rule Learning

Fürnkranz, Johannes; Kliegr, Tomáš

doi:10.1007/978-3-319-21542-6_4

Cited by 52 publications

(39 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To keep the benefit of an intelligible system with easily traceable decisions, rule learning or decision trees are a good candidate (Flach 2012). A recent overview of rule learning is published in (Fürnkranz and Kliegr 2015). By keeping the intermediate results of executions and tracking the execution time of each agent, this approach could be taken further to learn and incrementally improve the rule set online.…”

Section: Discussionmentioning

confidence: 99%

Supporting the building design process with graph-based methods using centrally coordinated federated databases

2017

View full text Add to dashboard Cite

Background: Technological developments and globalized working processes have transformed the building process. However, current digital semantic building models no longer adequately represent the increasing complexity of modern building projects. The potential of combining agent-based and graph-based methods, for example for energy calculations or spatial research strategies, is not fully exploited. Methods: In our system, users search for building floorplans, for example as sources of inspiration, by creating conceptual hand-drawn sketches of building parts on multi-touch devices. The sketch is analyzed and used to query a federated database system comprising a building information model server, the graph database neo4j and the content management system mediaTUM. Users interact with client applications that show and continuously update a list of floorplans by sending queries to a central coordinator service. In the paper, we describe the coordinator that enables our databases to appear as a single smart information system to search for digital information about buildings and visualize their floorplans. The application case comprises search by drawing the initial design idea of a building to find similar floorplans e.g. as source of inspiration. Results: Our federated database system is queried using semantic building floorplan fingerprints, which are formalized as graphs and encoded in a common schema, like our AGraphML, to represent spatial configurations and perform graph matching. As graph matching is computationally expensive, the coordinator needs to analyze the queries, separate different fingerprints and metadata and pass it to specialized agents, implemented as microservices, for processing. The returned result sets are combined and the results are visualized. The use of multiple agents facilitates the recombination of the data from the underlying disparate data sources and also serves to support different execution strategies which are chosen using properties of the query and a set of predefined rules. Conclusions: To deal with complex and dynamic queries, as well as dynamic results that are updated while some agents are still executing, a caching framework was developed which takes the similarity measures of the graph-based building representation into account when determining query equality.

show abstract

Section: Discussionmentioning

confidence: 99%

Supporting the building design process with graph-based methods using centrally coordinated federated databases

2017

View full text Add to dashboard Cite

show abstract

“…Hence, another method of significance is 'rules-based machine learning', which is a family of AI techniques that is based on application of a defined knowledge base, formulated as 'if … then …' rules. 34 For example, 'IF stock price moves X% AND currency moves Y% THEN buy'.…”

Section: Distinctive Features Of Deep Neural Network Versus 'Old-stymentioning

confidence: 99%

In Machines We Trust: Are Robo-Advisers More Trustworthy Than Human Financial Advisers?

Chia

2019

Law Tech Hum

View full text Add to dashboard Cite

show abstract

“…In this paper, we are interested in descriptive (or association) rules, for which there are three popular approaches: inductive logic programming, which uses logic programming as a uniform representation for examples, background knowledge and hypotheses, and aims at deriving a hypothesised logic program (that is, a set of rules) which entails all the positive and none of the negative examples (see, e.g., [18,19]); rule induction via metaheuristics, typically driven by evolutionary algorithms; and APRIORI [1] and its subsequent developments. These approaches have been extensively compared in the literature (see, e.g., [9] and references therein); apparently, although APRIORI is probably the first technology for rule extraction that gained some acknowledgment in the community, its main ideas are still widely used, since no negative examples are needed (in contrast to inductive logic programming), and since it is considered reliable and fast (in contrast to metaheuristic approaches, which are computationally expensive).…”

Section: Introductionmentioning

confidence: 99%

Towards a General Method for Logical Rule Extraction from Time Series

Sciavicco

Stan

Vaccari

2019

From Bioinspired Systems and Biomedical Applications to Machine Learning

View full text Add to dashboard Cite

Extracting rules from temporal series is a well-established temporal data mining technique. The current literature contains a number of different algorithms and experiments that allow one to abstract temporal series and, later, extract meaningful rules from them. In this paper, we approach this problem in a rather general way, without resorting, as many other methods, to expert knowledge and ad-hoc solutions. Our very simple temporal abstraction method allows us to transform time series into timelines, which can be then used for logical temporal rule extraction using an already existing temporal adaptation of the algorithm APRIORI. We have tested this approach on real data, obtaining promising results.

show abstract

A Brief Overview of Rule Learning

Cited by 52 publications

References 42 publications

Supporting the building design process with graph-based methods using centrally coordinated federated databases

Supporting the building design process with graph-based methods using centrally coordinated federated databases

In Machines We Trust: Are Robo-Advisers More Trustworthy Than Human Financial Advisers?

Towards a General Method for Logical Rule Extraction from Time Series

Contact Info

Product

Resources

About